Spacy - NER- Individual Data Processing + Model Training- Version 1.1.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# # Download \n",
    "\n",
    "# pip install -U spacy\n",
    "\n",
    "# !python -m spacy download en"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Training data structure for Sapcy\n",
    "```\n",
    "dummy_train_data = [\n",
    "    (\"Uber blew through $1 million a week\", [(0, 4, 'ORG')]),\n",
    "    (\"Android Pay expands to Canada\", [(0, 11, 'PRODUCT'), (23, 30, 'GPE')]),\n",
    "    (\"Spotify steps up Asia expansion\", [(0, 8, \"ORG\"), (17, 21, \"LOC\")]),\n",
    "    (\"Google Maps launches location sharing\", [(0, 11, \"PRODUCT\")]),\n",
    "    (\"Google rebrands its business apps\", [(0, 6, \"ORG\")]),\n",
    "    (\"look what i found on google! 😂\", [(21, 27, \"PRODUCT\")])]\n",
    "\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "filename = \"/home/ubuntu/Grakn/Spacy Play/pubmed_result_abstract-annotated_individual.json\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# with open(filename, encoding='utf-8') as train_data:\n",
    "# \ttrain = json.load(train_data)\n",
    "    \n",
    "data = []\n",
    "with open(filename) as f:\n",
    "    for line in f:\n",
    "        data.append(json.loads(line))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- data is a list containg dictionary of annotated text \n",
    "- It is a list of dictionary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'id': 2,\n",
       " 'text': 'The BRAF inhibitor dabrafenib (Tafinlar(Â®)) and the MEK inhibitor trametinib (Mekinist(Â®)) are indicated, as monotherapy or in combination with each other, for the treatment of patients with unresectable or metastatic melanoma with a BRAF (V600) mutation.',\n",
       " 'annotations': [{'label': 7, 'start_offset': 4, 'end_offset': 8, 'user': 1},\n",
       "  {'label': 8, 'start_offset': 19, 'end_offset': 29, 'user': 1},\n",
       "  {'label': 9, 'start_offset': 53, 'end_offset': 56, 'user': 1},\n",
       "  {'label': 8, 'start_offset': 67, 'end_offset': 77, 'user': 1},\n",
       "  {'label': 10, 'start_offset': 220, 'end_offset': 228, 'user': 1},\n",
       "  {'label': 7, 'start_offset': 236, 'end_offset': 240, 'user': 1},\n",
       "  {'label': 11, 'start_offset': 242, 'end_offset': 246, 'user': 1}],\n",
       " 'meta': {}}"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Dictionary for label used while annotation\n",
    "\n",
    "\n",
    "label_dic = {'7':'gene','8':'drug','9':'protien','10':'diesease','11':'varient'}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# for k,v in label_dic.items():\n",
    "#     print(k,v)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Iterating through all the dictionary of the list of dictionary of annotations\n",
    "\n",
    "for individual_dictionary in data:\n",
    "    for dic in individual_dictionary['annotations']:\n",
    "        # Going to find the label in the dictionary\n",
    "        temp_key = str(dic['label'])\n",
    "        #print(label_dic[temp_key])\n",
    "        dic['label'] = label_dic[temp_key]\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{'id': 2,\n",
       "  'text': 'The BRAF inhibitor dabrafenib (Tafinlar(Â®)) and the MEK inhibitor trametinib (Mekinist(Â®)) are indicated, as monotherapy or in combination with each other, for the treatment of patients with unresectable or metastatic melanoma with a BRAF (V600) mutation.',\n",
       "  'annotations': [{'label': 'gene',\n",
       "    'start_offset': 4,\n",
       "    'end_offset': 8,\n",
       "    'user': 1},\n",
       "   {'label': 'drug', 'start_offset': 19, 'end_offset': 29, 'user': 1},\n",
       "   {'label': 'protien', 'start_offset': 53, 'end_offset': 56, 'user': 1},\n",
       "   {'label': 'drug', 'start_offset': 67, 'end_offset': 77, 'user': 1},\n",
       "   {'label': 'diesease', 'start_offset': 220, 'end_offset': 228, 'user': 1},\n",
       "   {'label': 'gene', 'start_offset': 236, 'end_offset': 240, 'user': 1},\n",
       "   {'label': 'varient', 'start_offset': 242, 'end_offset': 246, 'user': 1}],\n",
       "  'meta': {}},\n",
       " {'id': 4,\n",
       "  'text': 'Dabrafenib plus trametinib significantly prolonged progression-free survival (PFS) and overall survival (OS), improved objective response rates (ORRs) and preserved health-related quality of life (HR-QOL) to a greater extent than dabrafenib (in the double-blind COMBI-d study) and vemurafenib (in the open-label COMBI-v study) in two large, randomized, phase III studies in treatment-naÃ¯ve patients with unresectable or metastatic melanoma with BRAF (V600E/K) mutation.',\n",
       "  'annotations': [{'label': 'drug',\n",
       "    'start_offset': 0,\n",
       "    'end_offset': 10,\n",
       "    'user': 1},\n",
       "   {'label': 'drug', 'start_offset': 16, 'end_offset': 26, 'user': 1},\n",
       "   {'label': 'drug', 'start_offset': 230, 'end_offset': 241, 'user': 1},\n",
       "   {'label': 'diesease', 'start_offset': 432, 'end_offset': 441, 'user': 1},\n",
       "   {'label': 'gene', 'start_offset': 446, 'end_offset': 450, 'user': 1},\n",
       "   {'label': 'varient', 'start_offset': 452, 'end_offset': 459, 'user': 1}],\n",
       "  'meta': {}},\n",
       " {'id': 5,\n",
       "  'text': 'Limited treatment benefit with the combination was also seen in patients who had progressed on prior BRAF inhibitor therapy, as indicated by ORRs of â‰¤ 15Â % and stable disease in â‰¤ 50Â % of patients in small phase I and II studies.Limited treatment benefit with the combination was also seen in patients who had progressed on prior BRAF inhibitor therapy, as indicated by ORRs of â‰¤ 15Â % and stable disease in â‰¤ 50Â % of patients in small phase I and II studies.',\n",
       "  'annotations': [{'label': 'gene',\n",
       "    'start_offset': 101,\n",
       "    'end_offset': 106,\n",
       "    'user': 1},\n",
       "   {'label': 'gene', 'start_offset': 336, 'end_offset': 341, 'user': 1}],\n",
       "  'meta': {}},\n",
       " {'id': 6,\n",
       "  'text': 'Combination therapy did not increase overall toxicity relative to dabrafenib or vemurafenib monotherapy, with most adverse events (AEs) mild or moderate in severity and generally manageable.',\n",
       "  'annotations': [{'label': 'drug',\n",
       "    'start_offset': 66,\n",
       "    'end_offset': 77,\n",
       "    'user': 1},\n",
       "   {'label': 'drug', 'start_offset': 80, 'end_offset': 92, 'user': 1}],\n",
       "  'meta': {}},\n",
       " {'id': 9,\n",
       "  'text': 'This article reviews the therapeutic efficacy and tolerability of combination treatment with dabrafenib and trametinib in this indication and summarizes relevant pharmacological data.',\n",
       "  'annotations': [{'label': 'drug',\n",
       "    'start_offset': 93,\n",
       "    'end_offset': 104,\n",
       "    'user': 1},\n",
       "   {'label': 'drug', 'start_offset': 108, 'end_offset': 119, 'user': 1}],\n",
       "  'meta': {}},\n",
       " {'id': 10,\n",
       "  'text': 'Fewer skin-related AEs (e.g. cutaneous malignancies, hyperkeratinosis and hand-foot syndrome) were reported with combination therapy than with dabrafenib or vemurafenib, probably because of reduced paradoxical activation of the MAPK pathway.',\n",
       "  'annotations': [{'label': 'drug',\n",
       "    'start_offset': 143,\n",
       "    'end_offset': 154,\n",
       "    'user': 1},\n",
       "   {'label': 'drug', 'start_offset': 157, 'end_offset': 168, 'user': 1}],\n",
       "  'meta': {}},\n",
       " {'id': 11,\n",
       "  'text': 'Thus, dabrafenib plus trametinib provides an important treatment option for patients with BRAF (V600) mutation-positive unresectable or metastatic melanoma.',\n",
       "  'annotations': [{'label': 'drug',\n",
       "    'start_offset': 6,\n",
       "    'end_offset': 17,\n",
       "    'user': 1},\n",
       "   {'label': 'drug', 'start_offset': 22, 'end_offset': 33, 'user': 1},\n",
       "   {'label': 'gene', 'start_offset': 90, 'end_offset': 95, 'user': 1},\n",
       "   {'label': 'varient', 'start_offset': 96, 'end_offset': 100, 'user': 1},\n",
       "   {'label': 'diesease', 'start_offset': 147, 'end_offset': 155, 'user': 1}],\n",
       "  'meta': {}}]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creating data lsit as per Spacy format\n",
    "# Iterating through all the dictionary for transforming in spacy data\n",
    "\n",
    "\n",
    "temp_data= []\n",
    "for d in data:\n",
    "    #print(d[\"text\"]) # print text of all dictionary\n",
    "    #print(d.keys()) # print keys of dictionary\n",
    "    temp_annotation = d['annotations']\n",
    "    temp_list = [] # for storing tuples\n",
    "    temp_data_list = [] # for storing list of tuples\n",
    "    for label_dic in temp_annotation:\n",
    "        temp_list.append((label_dic['start_offset'],label_dic['end_offset'],label_dic['label']))\n",
    "        #emp_list.append()\n",
    "        #emp_list.append()\n",
    "        #t(label_dic['start_offset'],label_dic['end_offset'],label_dic['label'])\n",
    "    temp_data.append((d['text'],{\"entities\":temp_list}))\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('The BRAF inhibitor dabrafenib (Tafinlar(Â®)) and the MEK inhibitor trametinib (Mekinist(Â®)) are indicated, as monotherapy or in combination with each other, for the treatment of patients with unresectable or metastatic melanoma with a BRAF (V600) mutation.',\n",
       "  {'entities': [(4, 8, 'gene'),\n",
       "    (19, 29, 'drug'),\n",
       "    (53, 56, 'protien'),\n",
       "    (67, 77, 'drug'),\n",
       "    (220, 228, 'diesease'),\n",
       "    (236, 240, 'gene'),\n",
       "    (242, 246, 'varient')]}),\n",
       " ('Dabrafenib plus trametinib significantly prolonged progression-free survival (PFS) and overall survival (OS), improved objective response rates (ORRs) and preserved health-related quality of life (HR-QOL) to a greater extent than dabrafenib (in the double-blind COMBI-d study) and vemurafenib (in the open-label COMBI-v study) in two large, randomized, phase III studies in treatment-naÃ¯ve patients with unresectable or metastatic melanoma with BRAF (V600E/K) mutation.',\n",
       "  {'entities': [(0, 10, 'drug'),\n",
       "    (16, 26, 'drug'),\n",
       "    (230, 241, 'drug'),\n",
       "    (432, 441, 'diesease'),\n",
       "    (446, 450, 'gene'),\n",
       "    (452, 459, 'varient')]}),\n",
       " ('Limited treatment benefit with the combination was also seen in patients who had progressed on prior BRAF inhibitor therapy, as indicated by ORRs of â‰¤ 15Â % and stable disease in â‰¤ 50Â % of patients in small phase I and II studies.Limited treatment benefit with the combination was also seen in patients who had progressed on prior BRAF inhibitor therapy, as indicated by ORRs of â‰¤ 15Â % and stable disease in â‰¤ 50Â % of patients in small phase I and II studies.',\n",
       "  {'entities': [(101, 106, 'gene'), (336, 341, 'gene')]}),\n",
       " ('Combination therapy did not increase overall toxicity relative to dabrafenib or vemurafenib monotherapy, with most adverse events (AEs) mild or moderate in severity and generally manageable.',\n",
       "  {'entities': [(66, 77, 'drug'), (80, 92, 'drug')]}),\n",
       " ('This article reviews the therapeutic efficacy and tolerability of combination treatment with dabrafenib and trametinib in this indication and summarizes relevant pharmacological data.',\n",
       "  {'entities': [(93, 104, 'drug'), (108, 119, 'drug')]}),\n",
       " ('Fewer skin-related AEs (e.g. cutaneous malignancies, hyperkeratinosis and hand-foot syndrome) were reported with combination therapy than with dabrafenib or vemurafenib, probably because of reduced paradoxical activation of the MAPK pathway.',\n",
       "  {'entities': [(143, 154, 'drug'), (157, 168, 'drug')]}),\n",
       " ('Thus, dabrafenib plus trametinib provides an important treatment option for patients with BRAF (V600) mutation-positive unresectable or metastatic melanoma.',\n",
       "  {'entities': [(6, 17, 'drug'),\n",
       "    (22, 33, 'drug'),\n",
       "    (90, 95, 'gene'),\n",
       "    (96, 100, 'varient'),\n",
       "    (147, 155, 'diesease')]})]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "temp_data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Preparing training data as per Spacy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "TRAIN_DATA = temp_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('The BRAF inhibitor dabrafenib (Tafinlar(Â®)) and the MEK inhibitor trametinib (Mekinist(Â®)) are indicated, as monotherapy or in combination with each other, for the treatment of patients with unresectable or metastatic melanoma with a BRAF (V600) mutation.',\n",
       "  {'entities': [(4, 8, 'gene'),\n",
       "    (19, 29, 'drug'),\n",
       "    (53, 56, 'protien'),\n",
       "    (67, 77, 'drug'),\n",
       "    (220, 228, 'diesease'),\n",
       "    (236, 240, 'gene'),\n",
       "    (242, 246, 'varient')]}),\n",
       " ('Dabrafenib plus trametinib significantly prolonged progression-free survival (PFS) and overall survival (OS), improved objective response rates (ORRs) and preserved health-related quality of life (HR-QOL) to a greater extent than dabrafenib (in the double-blind COMBI-d study) and vemurafenib (in the open-label COMBI-v study) in two large, randomized, phase III studies in treatment-naÃ¯ve patients with unresectable or metastatic melanoma with BRAF (V600E/K) mutation.',\n",
       "  {'entities': [(0, 10, 'drug'),\n",
       "    (16, 26, 'drug'),\n",
       "    (230, 241, 'drug'),\n",
       "    (432, 441, 'diesease'),\n",
       "    (446, 450, 'gene'),\n",
       "    (452, 459, 'varient')]}),\n",
       " ('Limited treatment benefit with the combination was also seen in patients who had progressed on prior BRAF inhibitor therapy, as indicated by ORRs of â‰¤ 15Â % and stable disease in â‰¤ 50Â % of patients in small phase I and II studies.Limited treatment benefit with the combination was also seen in patients who had progressed on prior BRAF inhibitor therapy, as indicated by ORRs of â‰¤ 15Â % and stable disease in â‰¤ 50Â % of patients in small phase I and II studies.',\n",
       "  {'entities': [(101, 106, 'gene'), (336, 341, 'gene')]}),\n",
       " ('Combination therapy did not increase overall toxicity relative to dabrafenib or vemurafenib monotherapy, with most adverse events (AEs) mild or moderate in severity and generally manageable.',\n",
       "  {'entities': [(66, 77, 'drug'), (80, 92, 'drug')]}),\n",
       " ('This article reviews the therapeutic efficacy and tolerability of combination treatment with dabrafenib and trametinib in this indication and summarizes relevant pharmacological data.',\n",
       "  {'entities': [(93, 104, 'drug'), (108, 119, 'drug')]}),\n",
       " ('Fewer skin-related AEs (e.g. cutaneous malignancies, hyperkeratinosis and hand-foot syndrome) were reported with combination therapy than with dabrafenib or vemurafenib, probably because of reduced paradoxical activation of the MAPK pathway.',\n",
       "  {'entities': [(143, 154, 'drug'), (157, 168, 'drug')]}),\n",
       " ('Thus, dabrafenib plus trametinib provides an important treatment option for patients with BRAF (V600) mutation-positive unresectable or metastatic melanoma.',\n",
       "  {'entities': [(6, 17, 'drug'),\n",
       "    (22, 33, 'drug'),\n",
       "    (90, 95, 'gene'),\n",
       "    (96, 100, 'varient'),\n",
       "    (147, 155, 'diesease')]})]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "TRAIN_DATA"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#  -- Training Model --"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "from __future__ import unicode_literals, print_function\n",
    "\n",
    "import plac\n",
    "import random\n",
    "from pathlib import Path\n",
    "import spacy\n",
    "from spacy.util import minibatch, compounding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "path_model_custom = \"/home/ubuntu/Grakn/Spacy Play/spacy_en_core_web_lg_V1.1\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/home/ubuntu/Grakn/Spacy Play/spacy_en_core_web_lg_V1.1'"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "path_model_custom"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# @plac.annotations(\n",
    "# model=(\"en\", \"option\", \"m\", str),\n",
    "# output_dir=(\"Optional output directory\", \"option\", \"o\", Path),\n",
    "# n_iter=(\"5\", \"option\", \"n\", int)\n",
    "#     )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# @plac.annotations(\n",
    "#     model=(\"Model name. Defaults to blank 'en' model.\", \"option\", \"m\", str),\n",
    "#     new_model_name=(\"New model name for model meta.\", \"option\", \"nm\", str),\n",
    "#     output_dir=(\"Optional output directory\", \"option\", \"o\", Path),\n",
    "#     n_iter=(\"Number of training iterations\", \"option\", \"n\", int),)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "def main(model=None, output_dir=None, n_iter=5000):\n",
    "    \"\"\"Load the model, set up the pipeline and train the entity recognizer.\"\"\"\n",
    "    if model is not None:\n",
    "        nlp = spacy.load(model)  # load existing spaCy model\n",
    "        print(\"Loaded model '%s'\" % model)\n",
    "    else:\n",
    "        nlp = spacy.blank(\"en\")  # create blank Language class\n",
    "        print(\"Created blank 'en' model\")\n",
    "\n",
    "    # create the built-in pipeline components and add them to the pipeline\n",
    "    # nlp.create_pipe works for built-ins that are registered with spaCy\n",
    "    if \"ner\" not in nlp.pipe_names:\n",
    "        ner = nlp.create_pipe(\"ner\")\n",
    "        nlp.add_pipe(ner, last=True)\n",
    "    # otherwise, get it so we can add labels\n",
    "    else:\n",
    "        ner = nlp.get_pipe(\"ner\")\n",
    "\n",
    "    # add labels\n",
    "    for _, annotations in TRAIN_DATA:\n",
    "        for ent in annotations.get(\"entities\"):\n",
    "            ner.add_label(ent[2])\n",
    "\n",
    "    # get names of other pipes to disable them during training\n",
    "    other_pipes = [pipe for pipe in nlp.pipe_names if pipe != \"ner\"]\n",
    "    with nlp.disable_pipes(*other_pipes):  # only train NER\n",
    "        # reset and initialize the weights randomly – but only if we're\n",
    "        # training a new model\n",
    "        if model is None:\n",
    "            nlp.begin_training()\n",
    "        for itn in range(n_iter):\n",
    "            random.shuffle(TRAIN_DATA)\n",
    "            losses = {}\n",
    "            # batch up the examples using spaCy's minibatch\n",
    "            batches = minibatch(TRAIN_DATA, size=compounding(4.0, 32.0, 1.001))\n",
    "            for batch in batches:\n",
    "                texts, annotations = zip(*batch)\n",
    "                nlp.update(\n",
    "                    texts,  # batch of texts\n",
    "                    annotations,  # batch of annotations\n",
    "                    drop=0.5,  # dropout - make it harder to memorise data\n",
    "                    losses=losses,\n",
    "                )\n",
    "            print(\"Losses\", losses)\n",
    "\n",
    "    # test the trained model\n",
    "    for text, _ in TRAIN_DATA:\n",
    "        doc = nlp(text)\n",
    "        print(\"Entities\", [(ent.text, ent.label_) for ent in doc.ents])\n",
    "        print(\"Tokens\", [(t.text, t.ent_type_, t.ent_iob) for t in doc])\n",
    "\n",
    "    # save model to output directory\n",
    "    if output_dir is not None:\n",
    "        output_dir = Path(output_dir)\n",
    "        if not output_dir.exists():\n",
    "            output_dir.mkdir()\n",
    "        nlp.to_disk(output_dir)\n",
    "        print(\"Saved model to\", output_dir)\n",
    "\n",
    "        # test the saved model\n",
    "        print(\"Loading from\", output_dir)\n",
    "        nlp2 = spacy.load(output_dir)\n",
    "        for text, _ in TRAIN_DATA:\n",
    "            doc = nlp2(text)\n",
    "            print(\"Entities\", [(ent.text, ent.label_) for ent in doc.ents])\n",
    "            print(\"Tokens\", [(t.text, t.ent_type_, t.ent_iob) for t in doc])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loaded model 'en_core_web_lg'\n",
      "Losses {'ner': 352.8377512693405}\n",
      "Losses {'ner': 350.4965937137604}\n",
      "Losses {'ner': 347.09135019779205}\n",
      "Losses {'ner': 339.43800961971283}\n",
      "Losses {'ner': 334.1063941717148}\n",
      "Losses {'ner': 329.49082386493683}\n",
      "Losses {'ner': 337.7781318426132}\n",
      "Losses {'ner': 315.86620676517487}\n",
      "Losses {'ner': 300.0443922281265}\n",
      "Losses {'ner': 317.15666550397873}\n",
      "Losses {'ner': 327.6762192249298}\n",
      "Losses {'ner': 306.8059129714966}\n",
      "Losses {'ner': 312.67649102211}\n",
      "Losses {'ner': 274.8668670654297}\n",
      "Losses {'ner': 305.30070304870605}\n",
      "Losses {'ner': 316.38429617881775}\n",
      "Losses {'ner': 298.81794142723083}\n",
      "Losses {'ner': 302.0459725856781}\n",
      "Losses {'ner': 290.71587014198303}\n",
      "Losses {'ner': 285.59963393211365}\n",
      "Losses {'ner': 298.9365756511688}\n",
      "Losses {'ner': 300.10131001472473}\n",
      "Losses {'ner': 286.76170778274536}\n",
      "Losses {'ner': 289.78033208847046}\n",
      "Losses {'ner': 282.8889833688736}\n",
      "Losses {'ner': 272.28366327285767}\n",
      "Losses {'ner': 270.30282402038574}\n",
      "Losses {'ner': 269.6124818325043}\n",
      "Losses {'ner': 279.27388095855713}\n",
      "Losses {'ner': 268.40699195861816}\n",
      "Losses {'ner': 285.93292593955994}\n",
      "Losses {'ner': 272.72224402427673}\n",
      "Losses {'ner': 267.4695394039154}\n",
      "Losses {'ner': 282.2676296234131}\n",
      "Losses {'ner': 248.91399216651917}\n",
      "Losses {'ner': 268.2114222049713}\n",
      "Losses {'ner': 271.06090891361237}\n",
      "Losses {'ner': 282.24003291130066}\n",
      "Losses {'ner': 271.508088350296}\n",
      "Losses {'ner': 265.484915971756}\n",
      "Losses {'ner': 258.1741580963135}\n",
      "Losses {'ner': 262.493980884552}\n",
      "Losses {'ner': 249.20045220851898}\n",
      "Losses {'ner': 258.2413305044174}\n",
      "Losses {'ner': 275.83066976070404}\n",
      "Losses {'ner': 263.5431845188141}\n",
      "Losses {'ner': 274.9402811527252}\n",
      "Losses {'ner': 277.2979805469513}\n",
      "Losses {'ner': 241.2334280014038}\n",
      "Losses {'ner': 263.58803963661194}\n",
      "Losses {'ner': 265.50761955976486}\n",
      "Losses {'ner': 272.7685286998749}\n",
      "Losses {'ner': 248.91285395622253}\n",
      "Losses {'ner': 263.32724261283875}\n",
      "Losses {'ner': 253.37748336791992}\n",
      "Losses {'ner': 268.3311104774475}\n",
      "Losses {'ner': 266.94580841064453}\n",
      "Losses {'ner': 249.56874907016754}\n",
      "Losses {'ner': 250.5729194879532}\n",
      "Losses {'ner': 270.1613347530365}\n",
      "Losses {'ner': 256.3307636976242}\n",
      "Losses {'ner': 275.68520855903625}\n",
      "Losses {'ner': 270.5681451559067}\n",
      "Losses {'ner': 261.33883833885193}\n",
      "Losses {'ner': 256.18726873397827}\n",
      "Losses {'ner': 252.6650391817093}\n",
      "Losses {'ner': 256.49968802928925}\n",
      "Losses {'ner': 248.45971703529358}\n",
      "Losses {'ner': 250.07728505134583}\n",
      "Losses {'ner': 261.40424489974976}\n",
      "Losses {'ner': 273.9574296474457}\n",
      "Losses {'ner': 249.00977087020874}\n",
      "Losses {'ner': 256.00757789611816}\n",
      "Losses {'ner': 249.98094296455383}\n",
      "Losses {'ner': 271.38078236579895}\n",
      "Losses {'ner': 251.60254001617432}\n",
      "Losses {'ner': 238.35753464698792}\n",
      "Losses {'ner': 257.01900804042816}\n",
      "Losses {'ner': 240.80312490463257}\n",
      "Losses {'ner': 230.8183925151825}\n",
      "Losses {'ner': 254.61678040027618}\n",
      "Losses {'ner': 249.76345372200012}\n",
      "Losses {'ner': 242.83121395111084}\n",
      "Losses {'ner': 227.04438638687134}\n",
      "Losses {'ner': 235.9163110256195}\n",
      "Losses {'ner': 247.69040629267693}\n",
      "Losses {'ner': 233.26126909255981}\n",
      "Losses {'ner': 236.24579095840454}\n",
      "Losses {'ner': 253.86378574371338}\n",
      "Losses {'ner': 247.78494668006897}\n",
      "Losses {'ner': 249.89305317401886}\n",
      "Losses {'ner': 256.171466588974}\n",
      "Losses {'ner': 256.4881236553192}\n",
      "Losses {'ner': 257.1439629793167}\n",
      "Losses {'ner': 249.2200047969818}\n",
      "Losses {'ner': 248.4841923713684}\n",
      "Losses {'ner': 238.29941248893738}\n",
      "Losses {'ner': 244.16045355796814}\n",
      "Losses {'ner': 253.66422641277313}\n",
      "Losses {'ner': 257.5129814147949}\n",
      "Losses {'ner': 250.76761853694916}\n",
      "Losses {'ner': 239.13087248802185}\n",
      "Losses {'ner': 240.36801171302795}\n",
      "Losses {'ner': 244.95548748970032}\n",
      "Losses {'ner': 244.48072290420532}\n",
      "Losses {'ner': 245.6826696395874}\n",
      "Losses {'ner': 265.44348526000977}\n",
      "Losses {'ner': 252.151731133461}\n",
      "Losses {'ner': 254.84995412826538}\n",
      "Losses {'ner': 239.7737603187561}\n",
      "Losses {'ner': 251.06258177757263}\n",
      "Losses {'ner': 237.9420205950737}\n",
      "Losses {'ner': 237.1056706905365}\n",
      "Losses {'ner': 248.46848344802856}\n",
      "Losses {'ner': 251.88652157783508}\n",
      "Losses {'ner': 232.24391412734985}\n",
      "Losses {'ner': 256.1627175807953}\n",
      "Losses {'ner': 248.44270300865173}\n",
      "Losses {'ner': 245.45029425621033}\n",
      "Losses {'ner': 237.73543214797974}\n",
      "Losses {'ner': 243.78439569473267}\n",
      "Losses {'ner': 253.5644826889038}\n",
      "Losses {'ner': 246.44382524490356}\n",
      "Losses {'ner': 258.77366375923157}\n",
      "Losses {'ner': 238.31798946857452}\n",
      "Losses {'ner': 236.66233658790588}\n",
      "Losses {'ner': 246.27966713905334}\n",
      "Losses {'ner': 227.4688966870308}\n",
      "Losses {'ner': 246.71433782577515}\n",
      "Losses {'ner': 231.19124388694763}\n",
      "Losses {'ner': 246.12221312522888}\n",
      "Losses {'ner': 238.80039846897125}\n",
      "Losses {'ner': 242.3747570514679}\n",
      "Losses {'ner': 253.54848623275757}\n",
      "Losses {'ner': 239.08141469955444}\n",
      "Losses {'ner': 229.1620362997055}\n",
      "Losses {'ner': 237.45037138462067}\n",
      "Losses {'ner': 247.7986171245575}\n",
      "Losses {'ner': 241.88535928726196}\n",
      "Losses {'ner': 260.4680524468422}\n",
      "Losses {'ner': 253.63452768325806}\n",
      "Losses {'ner': 233.08395910263062}\n",
      "Losses {'ner': 230.25341093540192}\n",
      "Losses {'ner': 233.92421102523804}\n",
      "Losses {'ner': 237.8552541732788}\n",
      "Losses {'ner': 246.0123714208603}\n",
      "Losses {'ner': 241.61088705062866}\n",
      "Losses {'ner': 238.59603834152222}\n",
      "Losses {'ner': 230.40066254138947}\n",
      "Losses {'ner': 234.95433139801025}\n",
      "Losses {'ner': 232.01585793495178}\n",
      "Losses {'ner': 244.34557831287384}\n",
      "Losses {'ner': 242.6980414390564}\n",
      "Losses {'ner': 224.0082403421402}\n",
      "Losses {'ner': 236.61727094650269}\n",
      "Losses {'ner': 237.03427267074585}\n",
      "Losses {'ner': 238.94848102331161}\n",
      "Losses {'ner': 240.93355250358582}\n",
      "Losses {'ner': 241.08260428905487}\n",
      "Losses {'ner': 236.40972435474396}\n",
      "Losses {'ner': 247.02651715278625}\n",
      "Losses {'ner': 237.99700319766998}\n",
      "Losses {'ner': 242.47227001190186}\n",
      "Losses {'ner': 240.57258439064026}\n",
      "Losses {'ner': 231.6441843509674}\n",
      "Losses {'ner': 251.896390914917}\n",
      "Losses {'ner': 238.8167549073696}\n",
      "Losses {'ner': 253.61303329467773}\n",
      "Losses {'ner': 248.51230025291443}\n",
      "Losses {'ner': 248.48295187950134}\n",
      "Losses {'ner': 238.23275995254517}\n",
      "Losses {'ner': 253.51534175872803}\n",
      "Losses {'ner': 235.71202111244202}\n",
      "Losses {'ner': 254.22944617271423}\n",
      "Losses {'ner': 219.42613461613655}\n",
      "Losses {'ner': 224.53929567337036}\n",
      "Losses {'ner': 238.62358260154724}\n",
      "Losses {'ner': 220.19272208213806}\n",
      "Losses {'ner': 241.62208497524261}\n",
      "Losses {'ner': 239.66530799865723}\n",
      "Losses {'ner': 244.41129279136658}\n",
      "Losses {'ner': 230.9082955121994}\n",
      "Losses {'ner': 242.72006356716156}\n",
      "Losses {'ner': 226.3792064189911}\n",
      "Losses {'ner': 235.9205048084259}\n",
      "Losses {'ner': 233.7055230140686}\n",
      "Losses {'ner': 226.52738440036774}\n",
      "Losses {'ner': 231.08666157722473}\n",
      "Losses {'ner': 244.36152637004852}\n",
      "Losses {'ner': 240.15378403663635}\n",
      "Losses {'ner': 239.80640196800232}\n",
      "Losses {'ner': 229.5127658843994}\n",
      "Losses {'ner': 238.88832437992096}\n",
      "Losses {'ner': 247.1640751361847}\n",
      "Losses {'ner': 247.09437894821167}\n",
      "Losses {'ner': 233.33334398269653}\n",
      "Losses {'ner': 248.83170664310455}\n",
      "Losses {'ner': 246.88771438598633}\n",
      "Losses {'ner': 239.1218158006668}\n",
      "Losses {'ner': 225.595547914505}\n",
      "Losses {'ner': 226.73658323287964}\n",
      "Losses {'ner': 243.40255188941956}\n",
      "Losses {'ner': 234.77842462062836}\n",
      "Losses {'ner': 233.1028652191162}\n",
      "Losses {'ner': 238.3456072807312}\n",
      "Losses {'ner': 233.7147696018219}\n",
      "Losses {'ner': 245.96613192558289}\n",
      "Losses {'ner': 249.89954090118408}\n",
      "Losses {'ner': 235.4086924791336}\n",
      "Losses {'ner': 224.13498973846436}\n",
      "Losses {'ner': 237.67663669586182}\n",
      "Losses {'ner': 239.20524895191193}\n",
      "Losses {'ner': 231.23214495182037}\n",
      "Losses {'ner': 234.68883728981018}\n",
      "Losses {'ner': 246.9341287612915}\n",
      "Losses {'ner': 236.46930027008057}\n",
      "Losses {'ner': 222.45758438110352}\n",
      "Losses {'ner': 238.21976816654205}\n",
      "Losses {'ner': 226.60895156860352}\n",
      "Losses {'ner': 230.57365226745605}\n",
      "Losses {'ner': 244.6611704826355}\n",
      "Losses {'ner': 243.52736806869507}\n",
      "Losses {'ner': 258.4422241449356}\n",
      "Losses {'ner': 230.5744730234146}\n",
      "Losses {'ner': 243.01088523864746}\n",
      "Losses {'ner': 240.54215598106384}\n",
      "Losses {'ner': 243.11848556995392}\n",
      "Losses {'ner': 246.29432368278503}\n",
      "Losses {'ner': 241.40805399417877}\n",
      "Losses {'ner': 243.33540987968445}\n",
      "Losses {'ner': 231.32247138023376}\n",
      "Losses {'ner': 236.37438559532166}\n",
      "Losses {'ner': 232.97547376155853}\n",
      "Losses {'ner': 248.14585828781128}\n",
      "Losses {'ner': 218.84254038333893}\n",
      "Losses {'ner': 245.89959526062012}\n",
      "Losses {'ner': 235.8190474510193}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 239.93462681770325}\n",
      "Losses {'ner': 241.19732069969177}\n",
      "Losses {'ner': 244.86050748825073}\n",
      "Losses {'ner': 237.88910913467407}\n",
      "Losses {'ner': 232.59402132034302}\n",
      "Losses {'ner': 232.69361817836761}\n",
      "Losses {'ner': 229.64554035663605}\n",
      "Losses {'ner': 235.84381294250488}\n",
      "Losses {'ner': 232.8939459323883}\n",
      "Losses {'ner': 236.98808300495148}\n",
      "Losses {'ner': 236.92073714733124}\n",
      "Losses {'ner': 240.1130223274231}\n",
      "Losses {'ner': 239.08926713466644}\n",
      "Losses {'ner': 236.44816398620605}\n",
      "Losses {'ner': 240.99107360839844}\n",
      "Losses {'ner': 231.82672595977783}\n",
      "Losses {'ner': 242.60791325569153}\n",
      "Losses {'ner': 248.28693521022797}\n",
      "Losses {'ner': 242.60981106758118}\n",
      "Losses {'ner': 232.48056972026825}\n",
      "Losses {'ner': 242.71939873695374}\n",
      "Losses {'ner': 226.67696177959442}\n",
      "Losses {'ner': 224.57269668579102}\n",
      "Losses {'ner': 238.7110915184021}\n",
      "Losses {'ner': 244.3305858373642}\n",
      "Losses {'ner': 244.14511036872864}\n",
      "Losses {'ner': 247.60634016990662}\n",
      "Losses {'ner': 242.53034687042236}\n",
      "Losses {'ner': 246.43781566619873}\n",
      "Losses {'ner': 240.50961327552795}\n",
      "Losses {'ner': 237.03406596183777}\n",
      "Losses {'ner': 224.7703915834427}\n",
      "Losses {'ner': 233.33631920814514}\n",
      "Losses {'ner': 240.8752624988556}\n",
      "Losses {'ner': 251.79783010482788}\n",
      "Losses {'ner': 240.4216115474701}\n",
      "Losses {'ner': 248.9087643623352}\n",
      "Losses {'ner': 241.833313703537}\n",
      "Losses {'ner': 229.5568528175354}\n",
      "Losses {'ner': 240.03906786441803}\n",
      "Losses {'ner': 235.5948867201805}\n",
      "Losses {'ner': 237.04894876480103}\n",
      "Losses {'ner': 244.16042578220367}\n",
      "Losses {'ner': 241.73452305793762}\n",
      "Losses {'ner': 240.72970116138458}\n",
      "Losses {'ner': 244.6167163848877}\n",
      "Losses {'ner': 238.71608138084412}\n",
      "Losses {'ner': 241.04548859596252}\n",
      "Losses {'ner': 233.36453223228455}\n",
      "Losses {'ner': 231.71555936336517}\n",
      "Losses {'ner': 232.14702677726746}\n",
      "Losses {'ner': 227.60498070716858}\n",
      "Losses {'ner': 223.11740851402283}\n",
      "Losses {'ner': 233.08389711380005}\n",
      "Losses {'ner': 245.24308133125305}\n",
      "Losses {'ner': 235.4669795036316}\n",
      "Losses {'ner': 223.55660462379456}\n",
      "Losses {'ner': 245.43341040611267}\n",
      "Losses {'ner': 244.01232194900513}\n",
      "Losses {'ner': 241.14105987548828}\n",
      "Losses {'ner': 235.85875248908997}\n",
      "Losses {'ner': 234.11932182312012}\n",
      "Losses {'ner': 228.02205538749695}\n",
      "Losses {'ner': 243.91920495033264}\n",
      "Losses {'ner': 224.31258130073547}\n",
      "Losses {'ner': 236.94067406654358}\n",
      "Losses {'ner': 223.02804517745972}\n",
      "Losses {'ner': 227.7321810722351}\n",
      "Losses {'ner': 233.52648639678955}\n",
      "Losses {'ner': 232.94423484802246}\n",
      "Losses {'ner': 231.50185841321945}\n",
      "Losses {'ner': 227.52474761009216}\n",
      "Losses {'ner': 240.6520323753357}\n",
      "Losses {'ner': 239.30288004875183}\n",
      "Losses {'ner': 244.99832272529602}\n",
      "Losses {'ner': 235.30843997001648}\n",
      "Losses {'ner': 227.64947605133057}\n",
      "Losses {'ner': 232.12916898727417}\n",
      "Losses {'ner': 233.63494956493378}\n",
      "Losses {'ner': 228.19016432762146}\n",
      "Losses {'ner': 239.88994765281677}\n",
      "Losses {'ner': 228.8738294839859}\n",
      "Losses {'ner': 235.45103286206722}\n",
      "Losses {'ner': 227.6535548567772}\n",
      "Losses {'ner': 217.41711473464966}\n",
      "Losses {'ner': 239.45343708992004}\n",
      "Losses {'ner': 250.69911921024323}\n",
      "Losses {'ner': 234.97493296861649}\n",
      "Losses {'ner': 234.37666988372803}\n",
      "Losses {'ner': 224.60896861553192}\n",
      "Losses {'ner': 231.73908519744873}\n",
      "Losses {'ner': 234.879070520401}\n",
      "Losses {'ner': 246.3942151069641}\n",
      "Losses {'ner': 237.6555676460266}\n",
      "Losses {'ner': 227.16933661699295}\n",
      "Losses {'ner': 249.74389624595642}\n",
      "Losses {'ner': 238.5441403388977}\n",
      "Losses {'ner': 239.03487348556519}\n",
      "Losses {'ner': 236.81882166862488}\n",
      "Losses {'ner': 234.425785779953}\n",
      "Losses {'ner': 232.29186010360718}\n",
      "Losses {'ner': 233.12675309181213}\n",
      "Losses {'ner': 233.65858960151672}\n",
      "Losses {'ner': 238.56725931167603}\n",
      "Losses {'ner': 229.91628336906433}\n",
      "Losses {'ner': 226.71347200870514}\n",
      "Losses {'ner': 230.73603177070618}\n",
      "Losses {'ner': 238.41235387325287}\n",
      "Losses {'ner': 251.14802622795105}\n",
      "Losses {'ner': 241.0267058610916}\n",
      "Losses {'ner': 238.9260847568512}\n",
      "Losses {'ner': 223.46761178970337}\n",
      "Losses {'ner': 233.65837860107422}\n",
      "Losses {'ner': 222.44019532203674}\n",
      "Losses {'ner': 236.44622802734375}\n",
      "Losses {'ner': 246.85479283332825}\n",
      "Losses {'ner': 243.96025693416595}\n",
      "Losses {'ner': 233.02579152584076}\n",
      "Losses {'ner': 230.31943225860596}\n",
      "Losses {'ner': 236.10393929481506}\n",
      "Losses {'ner': 226.58795261383057}\n",
      "Losses {'ner': 239.06265091896057}\n",
      "Losses {'ner': 229.53208756446838}\n",
      "Losses {'ner': 241.54449081420898}\n",
      "Losses {'ner': 224.84277379512787}\n",
      "Losses {'ner': 225.0253655910492}\n",
      "Losses {'ner': 231.46082758903503}\n",
      "Losses {'ner': 232.43431520462036}\n",
      "Losses {'ner': 226.40998244285583}\n",
      "Losses {'ner': 228.39400672912598}\n",
      "Losses {'ner': 236.1366674900055}\n",
      "Losses {'ner': 232.33186674118042}\n",
      "Losses {'ner': 240.897327542305}\n",
      "Losses {'ner': 242.04100728034973}\n",
      "Losses {'ner': 212.64366400241852}\n",
      "Losses {'ner': 235.5065597295761}\n",
      "Losses {'ner': 242.7513415813446}\n",
      "Losses {'ner': 242.49056720733643}\n",
      "Losses {'ner': 228.79834127426147}\n",
      "Losses {'ner': 237.13735604286194}\n",
      "Losses {'ner': 245.75387156009674}\n",
      "Losses {'ner': 236.23214530944824}\n",
      "Losses {'ner': 249.91267549991608}\n",
      "Losses {'ner': 236.04602909088135}\n",
      "Losses {'ner': 241.03423190116882}\n",
      "Losses {'ner': 241.24852299690247}\n",
      "Losses {'ner': 230.86219516396523}\n",
      "Losses {'ner': 236.2739099264145}\n",
      "Losses {'ner': 219.01155161857605}\n",
      "Losses {'ner': 229.7143862247467}\n",
      "Losses {'ner': 227.1210435628891}\n",
      "Losses {'ner': 230.20974016189575}\n",
      "Losses {'ner': 224.7642948627472}\n",
      "Losses {'ner': 233.2036213874817}\n",
      "Losses {'ner': 232.18059301376343}\n",
      "Losses {'ner': 234.79765072464943}\n",
      "Losses {'ner': 230.99074053764343}\n",
      "Losses {'ner': 239.62557435035706}\n",
      "Losses {'ner': 230.44375550746918}\n",
      "Losses {'ner': 223.5777187347412}\n",
      "Losses {'ner': 217.90202525258064}\n",
      "Losses {'ner': 228.81994426250458}\n",
      "Losses {'ner': 240.05927872657776}\n",
      "Losses {'ner': 226.9852772951126}\n",
      "Losses {'ner': 235.88007164001465}\n",
      "Losses {'ner': 225.92950010299683}\n",
      "Losses {'ner': 232.15826427936554}\n",
      "Losses {'ner': 238.3118156194687}\n",
      "Losses {'ner': 235.30592727661133}\n",
      "Losses {'ner': 237.065593957901}\n",
      "Losses {'ner': 238.74278938770294}\n",
      "Losses {'ner': 239.7880675792694}\n",
      "Losses {'ner': 232.27764344215393}\n",
      "Losses {'ner': 235.20116567611694}\n",
      "Losses {'ner': 231.83768701553345}\n",
      "Losses {'ner': 232.46558809280396}\n",
      "Losses {'ner': 233.6908757686615}\n",
      "Losses {'ner': 222.33389115333557}\n",
      "Losses {'ner': 231.79569625854492}\n",
      "Losses {'ner': 221.93952012062073}\n",
      "Losses {'ner': 237.0266368985176}\n",
      "Losses {'ner': 227.4896833896637}\n",
      "Losses {'ner': 246.11848735809326}\n",
      "Losses {'ner': 228.29835081100464}\n",
      "Losses {'ner': 226.75150609016418}\n",
      "Losses {'ner': 234.71256113052368}\n",
      "Losses {'ner': 234.34709334373474}\n",
      "Losses {'ner': 242.4941942691803}\n",
      "Losses {'ner': 228.83156597614288}\n",
      "Losses {'ner': 222.86917996406555}\n",
      "Losses {'ner': 233.05043196678162}\n",
      "Losses {'ner': 237.87392485141754}\n",
      "Losses {'ner': 223.1376931667328}\n",
      "Losses {'ner': 240.82184624671936}\n",
      "Losses {'ner': 242.03353929519653}\n",
      "Losses {'ner': 233.27431738376617}\n",
      "Losses {'ner': 233.51822686195374}\n",
      "Losses {'ner': 226.168701171875}\n",
      "Losses {'ner': 226.57409393787384}\n",
      "Losses {'ner': 242.60443592071533}\n",
      "Losses {'ner': 226.2550654411316}\n",
      "Losses {'ner': 238.43963611125946}\n",
      "Losses {'ner': 231.61779564619064}\n",
      "Losses {'ner': 238.3206547498703}\n",
      "Losses {'ner': 232.5639596581459}\n",
      "Losses {'ner': 228.15037369728088}\n",
      "Losses {'ner': 226.18025267124176}\n",
      "Losses {'ner': 228.6218023300171}\n",
      "Losses {'ner': 219.87583589553833}\n",
      "Losses {'ner': 230.71710121631622}\n",
      "Losses {'ner': 228.71888947486877}\n",
      "Losses {'ner': 219.31913936138153}\n",
      "Losses {'ner': 232.08379435539246}\n",
      "Losses {'ner': 227.9270340204239}\n",
      "Losses {'ner': 231.46896302700043}\n",
      "Losses {'ner': 235.6911370754242}\n",
      "Losses {'ner': 240.11784553527832}\n",
      "Losses {'ner': 222.1047661304474}\n",
      "Losses {'ner': 220.21666598320007}\n",
      "Losses {'ner': 229.9375238418579}\n",
      "Losses {'ner': 243.9358353614807}\n",
      "Losses {'ner': 234.7241506576538}\n",
      "Losses {'ner': 234.39785718917847}\n",
      "Losses {'ner': 225.6020543575287}\n",
      "Losses {'ner': 224.5229103565216}\n",
      "Losses {'ner': 232.16577339172363}\n",
      "Losses {'ner': 243.8582203388214}\n",
      "Losses {'ner': 236.711181640625}\n",
      "Losses {'ner': 236.3435037136078}\n",
      "Losses {'ner': 227.5481903553009}\n",
      "Losses {'ner': 241.9153928756714}\n",
      "Losses {'ner': 219.6469075679779}\n",
      "Losses {'ner': 242.5092670917511}\n",
      "Losses {'ner': 228.54971528053284}\n",
      "Losses {'ner': 228.76330304145813}\n",
      "Losses {'ner': 234.37903213500977}\n",
      "Losses {'ner': 240.95682406425476}\n",
      "Losses {'ner': 248.30397999286652}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 233.53319120407104}\n",
      "Losses {'ner': 250.98040437698364}\n",
      "Losses {'ner': 240.81524801254272}\n",
      "Losses {'ner': 223.14396572113037}\n",
      "Losses {'ner': 231.74489903450012}\n",
      "Losses {'ner': 240.21763253211975}\n",
      "Losses {'ner': 234.81889951229095}\n",
      "Losses {'ner': 235.18732917308807}\n",
      "Losses {'ner': 217.9712656736374}\n",
      "Losses {'ner': 241.39117217063904}\n",
      "Losses {'ner': 227.76579916477203}\n",
      "Losses {'ner': 234.01616764068604}\n",
      "Losses {'ner': 237.38059854507446}\n",
      "Losses {'ner': 227.95936393737793}\n",
      "Losses {'ner': 231.2662916779518}\n",
      "Losses {'ner': 227.721289396286}\n",
      "Losses {'ner': 232.08705806732178}\n",
      "Losses {'ner': 232.64385199546814}\n",
      "Losses {'ner': 228.61735701560974}\n",
      "Losses {'ner': 233.20625746250153}\n",
      "Losses {'ner': 239.89596128463745}\n",
      "Losses {'ner': 233.91680598258972}\n",
      "Losses {'ner': 240.89523661136627}\n",
      "Losses {'ner': 238.41408824920654}\n",
      "Losses {'ner': 228.21961665153503}\n",
      "Losses {'ner': 226.8336465358734}\n",
      "Losses {'ner': 240.81942737102509}\n",
      "Losses {'ner': 231.33884191513062}\n",
      "Losses {'ner': 232.12781071662903}\n",
      "Losses {'ner': 234.31606805324554}\n",
      "Losses {'ner': 238.36813163757324}\n",
      "Losses {'ner': 240.2802152633667}\n",
      "Losses {'ner': 227.03778111934662}\n",
      "Losses {'ner': 241.9847731590271}\n",
      "Losses {'ner': 238.88230729103088}\n",
      "Losses {'ner': 234.01663482189178}\n",
      "Losses {'ner': 236.45896911621094}\n",
      "Losses {'ner': 234.9748638868332}\n",
      "Losses {'ner': 224.15215349197388}\n",
      "Losses {'ner': 227.32494592666626}\n",
      "Losses {'ner': 227.28466498851776}\n",
      "Losses {'ner': 232.07360816001892}\n",
      "Losses {'ner': 221.96898287534714}\n",
      "Losses {'ner': 223.00594985485077}\n",
      "Losses {'ner': 228.8972897529602}\n",
      "Losses {'ner': 241.59576225280762}\n",
      "Losses {'ner': 238.84022045135498}\n",
      "Losses {'ner': 238.51977407932281}\n",
      "Losses {'ner': 232.41052198410034}\n",
      "Losses {'ner': 231.07651162147522}\n",
      "Losses {'ner': 231.08084678649902}\n",
      "Losses {'ner': 233.73579961061478}\n",
      "Losses {'ner': 224.03932881355286}\n",
      "Losses {'ner': 233.14967322349548}\n",
      "Losses {'ner': 237.88736009597778}\n",
      "Losses {'ner': 233.5605971813202}\n",
      "Losses {'ner': 231.1956894993782}\n",
      "Losses {'ner': 233.85510778427124}\n",
      "Losses {'ner': 221.9362713098526}\n",
      "Losses {'ner': 237.98282438516617}\n",
      "Losses {'ner': 226.39415168762207}\n",
      "Losses {'ner': 232.17656141519547}\n",
      "Losses {'ner': 220.48581218719482}\n",
      "Losses {'ner': 217.7240093946457}\n",
      "Losses {'ner': 235.8471176624298}\n",
      "Losses {'ner': 235.07930195331573}\n",
      "Losses {'ner': 222.06154370307922}\n",
      "Losses {'ner': 221.88618278503418}\n",
      "Losses {'ner': 237.55978846549988}\n",
      "Losses {'ner': 242.24220478534698}\n",
      "Losses {'ner': 237.6034529209137}\n",
      "Losses {'ner': 232.00685906410217}\n",
      "Losses {'ner': 222.0507528781891}\n",
      "Losses {'ner': 237.2962987422943}\n",
      "Losses {'ner': 241.38459300994873}\n",
      "Losses {'ner': 231.33609247207642}\n",
      "Losses {'ner': 233.45485925674438}\n",
      "Losses {'ner': 237.47049593925476}\n",
      "Losses {'ner': 224.29167640209198}\n",
      "Losses {'ner': 244.73494935035706}\n",
      "Losses {'ner': 229.97608470916748}\n",
      "Losses {'ner': 229.75508654117584}\n",
      "Losses {'ner': 224.24543917179108}\n",
      "Losses {'ner': 219.32308673858643}\n",
      "Losses {'ner': 227.2000030875206}\n",
      "Losses {'ner': 239.6685711145401}\n",
      "Losses {'ner': 230.9173400402069}\n",
      "Losses {'ner': 222.43638706207275}\n",
      "Losses {'ner': 224.0631104707718}\n",
      "Losses {'ner': 230.76112365722656}\n",
      "Losses {'ner': 238.75225281715393}\n",
      "Losses {'ner': 213.37944388389587}\n",
      "Losses {'ner': 233.25279831886292}\n",
      "Losses {'ner': 239.61999011039734}\n",
      "Losses {'ner': 238.2178030014038}\n",
      "Losses {'ner': 221.46919918060303}\n",
      "Losses {'ner': 229.53802490234375}\n",
      "Losses {'ner': 217.55821204185486}\n",
      "Losses {'ner': 221.9860508441925}\n",
      "Losses {'ner': 231.62541270256042}\n",
      "Losses {'ner': 219.29236352443695}\n",
      "Losses {'ner': 217.393967628479}\n",
      "Losses {'ner': 231.34863245487213}\n",
      "Losses {'ner': 248.51078152656555}\n",
      "Losses {'ner': 236.49507546424866}\n",
      "Losses {'ner': 249.6905950307846}\n",
      "Losses {'ner': 229.3830268383026}\n",
      "Losses {'ner': 228.1801837682724}\n",
      "Losses {'ner': 235.04958128929138}\n",
      "Losses {'ner': 232.34022867679596}\n",
      "Losses {'ner': 230.2938678264618}\n",
      "Losses {'ner': 237.98249912261963}\n",
      "Losses {'ner': 246.5268111228943}\n",
      "Losses {'ner': 221.8945506811142}\n",
      "Losses {'ner': 251.60751152038574}\n",
      "Losses {'ner': 230.65219902992249}\n",
      "Losses {'ner': 227.04497694969177}\n",
      "Losses {'ner': 228.89908111095428}\n",
      "Losses {'ner': 231.91889333724976}\n",
      "Losses {'ner': 226.80265307426453}\n",
      "Losses {'ner': 219.50854349136353}\n",
      "Losses {'ner': 233.69179391860962}\n",
      "Losses {'ner': 226.36089181900024}\n",
      "Losses {'ner': 228.11853086948395}\n",
      "Losses {'ner': 229.68107867240906}\n",
      "Losses {'ner': 223.83132827281952}\n",
      "Losses {'ner': 221.3059674501419}\n",
      "Losses {'ner': 225.68005430698395}\n",
      "Losses {'ner': 239.1672111749649}\n",
      "Losses {'ner': 228.607787668705}\n",
      "Losses {'ner': 215.3562434911728}\n",
      "Losses {'ner': 226.49625670909882}\n",
      "Losses {'ner': 225.68080854415894}\n",
      "Losses {'ner': 239.93317770957947}\n",
      "Losses {'ner': 223.80670964717865}\n",
      "Losses {'ner': 230.8024309873581}\n",
      "Losses {'ner': 228.5945382118225}\n",
      "Losses {'ner': 238.9782271385193}\n",
      "Losses {'ner': 227.6409022808075}\n",
      "Losses {'ner': 225.93046522140503}\n",
      "Losses {'ner': 221.41830015182495}\n",
      "Losses {'ner': 232.34103870391846}\n",
      "Losses {'ner': 234.57351171970367}\n",
      "Losses {'ner': 228.42176485061646}\n",
      "Losses {'ner': 233.50079464912415}\n",
      "Losses {'ner': 235.99275279045105}\n",
      "Losses {'ner': 240.50742959976196}\n",
      "Losses {'ner': 227.81900292634964}\n",
      "Losses {'ner': 223.84305381774902}\n",
      "Losses {'ner': 222.08942663669586}\n",
      "Losses {'ner': 225.62943363189697}\n",
      "Losses {'ner': 211.19296669960022}\n",
      "Losses {'ner': 232.62000799179077}\n",
      "Losses {'ner': 209.4382951259613}\n",
      "Losses {'ner': 237.89315855503082}\n",
      "Losses {'ner': 227.40233021974564}\n",
      "Losses {'ner': 226.02945852279663}\n",
      "Losses {'ner': 227.16403651237488}\n",
      "Losses {'ner': 229.22080099582672}\n",
      "Losses {'ner': 232.55651712417603}\n",
      "Losses {'ner': 224.46383380889893}\n",
      "Losses {'ner': 219.2925431728363}\n",
      "Losses {'ner': 231.16028022766113}\n",
      "Losses {'ner': 230.36677622795105}\n",
      "Losses {'ner': 205.19174659252167}\n",
      "Losses {'ner': 237.72659134864807}\n",
      "Losses {'ner': 223.99010229110718}\n",
      "Losses {'ner': 239.87384271621704}\n",
      "Losses {'ner': 238.38654494285583}\n",
      "Losses {'ner': 229.08786010742188}\n",
      "Losses {'ner': 226.4812934398651}\n",
      "Losses {'ner': 223.5428762435913}\n",
      "Losses {'ner': 231.02707731723785}\n",
      "Losses {'ner': 222.17375338077545}\n",
      "Losses {'ner': 234.26116466522217}\n",
      "Losses {'ner': 224.44630193710327}\n",
      "Losses {'ner': 216.98281908035278}\n",
      "Losses {'ner': 222.89494514465332}\n",
      "Losses {'ner': 221.81354749202728}\n",
      "Losses {'ner': 233.63322496414185}\n",
      "Losses {'ner': 244.78680896759033}\n",
      "Losses {'ner': 241.6314091682434}\n",
      "Losses {'ner': 220.2624728679657}\n",
      "Losses {'ner': 239.31784987449646}\n",
      "Losses {'ner': 231.5911898612976}\n",
      "Losses {'ner': 216.69481885433197}\n",
      "Losses {'ner': 234.37857329845428}\n",
      "Losses {'ner': 236.8466911315918}\n",
      "Losses {'ner': 236.77348852157593}\n",
      "Losses {'ner': 223.9633491039276}\n",
      "Losses {'ner': 228.0896954536438}\n",
      "Losses {'ner': 218.83596062660217}\n",
      "Losses {'ner': 229.3094072341919}\n",
      "Losses {'ner': 224.46698236465454}\n",
      "Losses {'ner': 226.57176005840302}\n",
      "Losses {'ner': 224.77710819244385}\n",
      "Losses {'ner': 228.80264735221863}\n",
      "Losses {'ner': 229.9265034198761}\n",
      "Losses {'ner': 227.42994076013565}\n",
      "Losses {'ner': 229.87920689582825}\n",
      "Losses {'ner': 219.87751388549805}\n",
      "Losses {'ner': 237.93260991573334}\n",
      "Losses {'ner': 229.0657969713211}\n",
      "Losses {'ner': 231.75206983089447}\n",
      "Losses {'ner': 229.81020760536194}\n",
      "Losses {'ner': 224.73970928788185}\n",
      "Losses {'ner': 229.14307641983032}\n",
      "Losses {'ner': 229.7964732646942}\n",
      "Losses {'ner': 234.4896218776703}\n",
      "Losses {'ner': 234.81848680973053}\n",
      "Losses {'ner': 222.47037625312805}\n",
      "Losses {'ner': 221.58020567893982}\n",
      "Losses {'ner': 219.89910972118378}\n",
      "Losses {'ner': 228.32909488677979}\n",
      "Losses {'ner': 231.32077383995056}\n",
      "Losses {'ner': 230.2561695575714}\n",
      "Losses {'ner': 229.01366710662842}\n",
      "Losses {'ner': 223.04757595062256}\n",
      "Losses {'ner': 230.6930923461914}\n",
      "Losses {'ner': 224.1273238658905}\n",
      "Losses {'ner': 229.15741354227066}\n",
      "Losses {'ner': 232.306809425354}\n",
      "Losses {'ner': 225.51109111309052}\n",
      "Losses {'ner': 223.04153060913086}\n",
      "Losses {'ner': 230.96716701984406}\n",
      "Losses {'ner': 241.60204339027405}\n",
      "Losses {'ner': 222.79746675491333}\n",
      "Losses {'ner': 230.25645327568054}\n",
      "Losses {'ner': 225.3011405467987}\n",
      "Losses {'ner': 234.39953589439392}\n",
      "Losses {'ner': 216.530611038208}\n",
      "Losses {'ner': 232.87788558006287}\n",
      "Losses {'ner': 224.60616254806519}\n",
      "Losses {'ner': 232.76741480827332}\n",
      "Losses {'ner': 239.37082588672638}\n",
      "Losses {'ner': 217.68685865402222}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 232.74148178100586}\n",
      "Losses {'ner': 230.06042650341988}\n",
      "Losses {'ner': 230.95778012275696}\n",
      "Losses {'ner': 235.75087106227875}\n",
      "Losses {'ner': 224.61944103240967}\n",
      "Losses {'ner': 228.98784351348877}\n",
      "Losses {'ner': 225.60899829864502}\n",
      "Losses {'ner': 235.98042726516724}\n",
      "Losses {'ner': 221.6649852991104}\n",
      "Losses {'ner': 232.39067149162292}\n",
      "Losses {'ner': 225.33304476737976}\n",
      "Losses {'ner': 224.92514967918396}\n",
      "Losses {'ner': 224.8810133934021}\n",
      "Losses {'ner': 221.9727818965912}\n",
      "Losses {'ner': 243.2869155406952}\n",
      "Losses {'ner': 233.2237401008606}\n",
      "Losses {'ner': 232.91923928260803}\n",
      "Losses {'ner': 226.50521397590637}\n",
      "Losses {'ner': 230.09517216682434}\n",
      "Losses {'ner': 225.48197388648987}\n",
      "Losses {'ner': 235.1260588169098}\n",
      "Losses {'ner': 236.46498572826385}\n",
      "Losses {'ner': 231.0814791917801}\n",
      "Losses {'ner': 230.6223304271698}\n",
      "Losses {'ner': 235.5310491323471}\n",
      "Losses {'ner': 234.47790050506592}\n",
      "Losses {'ner': 242.4646691083908}\n",
      "Losses {'ner': 219.95484828948975}\n",
      "Losses {'ner': 220.86176377534866}\n",
      "Losses {'ner': 215.71883046627045}\n",
      "Losses {'ner': 233.05308628082275}\n",
      "Losses {'ner': 227.85730719566345}\n",
      "Losses {'ner': 228.6361744403839}\n",
      "Losses {'ner': 234.84266304969788}\n",
      "Losses {'ner': 239.3316045999527}\n",
      "Losses {'ner': 217.1292942762375}\n",
      "Losses {'ner': 236.20882773399353}\n",
      "Losses {'ner': 234.4284520149231}\n",
      "Losses {'ner': 241.84336471557617}\n",
      "Losses {'ner': 227.12567389011383}\n",
      "Losses {'ner': 220.79031765460968}\n",
      "Losses {'ner': 228.7823600769043}\n",
      "Losses {'ner': 228.67986869812012}\n",
      "Losses {'ner': 224.4659389257431}\n",
      "Losses {'ner': 240.4615442752838}\n",
      "Losses {'ner': 223.44542706012726}\n",
      "Losses {'ner': 234.5176485478878}\n",
      "Losses {'ner': 237.06794786453247}\n",
      "Losses {'ner': 217.59408903121948}\n",
      "Losses {'ner': 227.55902314186096}\n",
      "Losses {'ner': 231.3051654100418}\n",
      "Losses {'ner': 224.94095611572266}\n",
      "Losses {'ner': 231.6721227169037}\n",
      "Losses {'ner': 223.28603172302246}\n",
      "Losses {'ner': 231.87431859970093}\n",
      "Losses {'ner': 217.22873878479004}\n",
      "Losses {'ner': 236.46496105194092}\n",
      "Losses {'ner': 223.63418233394623}\n",
      "Losses {'ner': 223.56230053305626}\n",
      "Losses {'ner': 240.82973456382751}\n",
      "Losses {'ner': 215.3684520125389}\n",
      "Losses {'ner': 220.45722138881683}\n",
      "Losses {'ner': 228.10620880126953}\n",
      "Losses {'ner': 228.85707545280457}\n",
      "Losses {'ner': 226.87801098823547}\n",
      "Losses {'ner': 231.92574858665466}\n",
      "Losses {'ner': 227.285768866539}\n",
      "Losses {'ner': 237.3954997062683}\n",
      "Losses {'ner': 238.5771745443344}\n",
      "Losses {'ner': 227.524352312088}\n",
      "Losses {'ner': 228.68828082084656}\n",
      "Losses {'ner': 223.94582748413086}\n",
      "Losses {'ner': 235.9013305902481}\n",
      "Losses {'ner': 229.15869963169098}\n",
      "Losses {'ner': 233.94567358493805}\n",
      "Losses {'ner': 230.7927601337433}\n",
      "Losses {'ner': 234.46737623214722}\n",
      "Losses {'ner': 219.10051095485687}\n",
      "Losses {'ner': 227.52329397201538}\n",
      "Losses {'ner': 224.70650959014893}\n",
      "Losses {'ner': 230.10359680652618}\n",
      "Losses {'ner': 222.88602089881897}\n",
      "Losses {'ner': 217.73981428146362}\n",
      "Losses {'ner': 222.7224222421646}\n",
      "Losses {'ner': 236.19793662428856}\n",
      "Losses {'ner': 232.51567459106445}\n",
      "Losses {'ner': 225.89103412628174}\n",
      "Losses {'ner': 222.01040443778038}\n",
      "Losses {'ner': 223.2865583896637}\n",
      "Losses {'ner': 232.19683957099915}\n",
      "Losses {'ner': 230.1699149608612}\n",
      "Losses {'ner': 221.7005696296692}\n",
      "Losses {'ner': 226.80208468437195}\n",
      "Losses {'ner': 225.66452956199646}\n",
      "Losses {'ner': 226.1911916732788}\n",
      "Losses {'ner': 223.36660194396973}\n",
      "Losses {'ner': 220.41422200202942}\n",
      "Losses {'ner': 235.04081165790558}\n",
      "Losses {'ner': 222.42978703975677}\n",
      "Losses {'ner': 239.73249077796936}\n",
      "Losses {'ner': 228.59740144014359}\n",
      "Losses {'ner': 249.2567223906517}\n",
      "Losses {'ner': 233.8837811946869}\n",
      "Losses {'ner': 234.44640243053436}\n",
      "Losses {'ner': 235.20540761947632}\n",
      "Losses {'ner': 215.20803093910217}\n",
      "Losses {'ner': 226.86029303073883}\n",
      "Losses {'ner': 228.66868710517883}\n",
      "Losses {'ner': 236.16402530670166}\n",
      "Losses {'ner': 247.39313507080078}\n",
      "Losses {'ner': 241.4282066822052}\n",
      "Losses {'ner': 221.28932285308838}\n",
      "Losses {'ner': 234.97034084796906}\n",
      "Losses {'ner': 227.50234246253967}\n",
      "Losses {'ner': 223.8645819425583}\n",
      "Losses {'ner': 237.20511627197266}\n",
      "Losses {'ner': 229.26966500282288}\n",
      "Losses {'ner': 207.89526963233948}\n",
      "Losses {'ner': 227.8996067047119}\n",
      "Losses {'ner': 222.30395317077637}\n",
      "Losses {'ner': 222.70995545387268}\n",
      "Losses {'ner': 236.31931900978088}\n",
      "Losses {'ner': 223.04375410079956}\n",
      "Losses {'ner': 222.32110059261322}\n",
      "Losses {'ner': 226.39372396469116}\n",
      "Losses {'ner': 236.81118392944336}\n",
      "Losses {'ner': 223.01068830490112}\n",
      "Losses {'ner': 209.3831639289856}\n",
      "Losses {'ner': 224.00549244880676}\n",
      "Losses {'ner': 228.36169159412384}\n",
      "Losses {'ner': 212.03009963035583}\n",
      "Losses {'ner': 232.78467416763306}\n",
      "Losses {'ner': 229.98405385017395}\n",
      "Losses {'ner': 223.45747864246368}\n",
      "Losses {'ner': 230.18174862861633}\n",
      "Losses {'ner': 233.82389187812805}\n",
      "Losses {'ner': 231.68164587020874}\n",
      "Losses {'ner': 226.49528002738953}\n",
      "Losses {'ner': 224.57506930828094}\n",
      "Losses {'ner': 226.66087114810944}\n",
      "Losses {'ner': 236.75526225566864}\n",
      "Losses {'ner': 222.6838936805725}\n",
      "Losses {'ner': 236.12932455539703}\n",
      "Losses {'ner': 222.64640319347382}\n",
      "Losses {'ner': 216.09270223975182}\n",
      "Losses {'ner': 230.06592512130737}\n",
      "Losses {'ner': 221.82329857349396}\n",
      "Losses {'ner': 231.95103430747986}\n",
      "Losses {'ner': 232.34714078903198}\n",
      "Losses {'ner': 216.60336327552795}\n",
      "Losses {'ner': 220.4210776090622}\n",
      "Losses {'ner': 220.67946767807007}\n",
      "Losses {'ner': 239.08062720298767}\n",
      "Losses {'ner': 220.55059957504272}\n",
      "Losses {'ner': 226.14484375715256}\n",
      "Losses {'ner': 236.7176194190979}\n",
      "Losses {'ner': 220.6628544330597}\n",
      "Losses {'ner': 227.3357870578766}\n",
      "Losses {'ner': 231.59351754188538}\n",
      "Losses {'ner': 218.6456274986267}\n",
      "Losses {'ner': 220.925390958786}\n",
      "Losses {'ner': 229.0036859512329}\n",
      "Losses {'ner': 233.00045680999756}\n",
      "Losses {'ner': 215.93701028823853}\n",
      "Losses {'ner': 211.7844636440277}\n",
      "Losses {'ner': 221.40780091285706}\n",
      "Losses {'ner': 222.3507609963417}\n",
      "Losses {'ner': 229.65426874160767}\n",
      "Losses {'ner': 221.471337556839}\n",
      "Losses {'ner': 215.05441546440125}\n",
      "Losses {'ner': 230.44552898406982}\n",
      "Losses {'ner': 225.04983687400818}\n",
      "Losses {'ner': 239.7209873199463}\n",
      "Losses {'ner': 219.85937249660492}\n",
      "Losses {'ner': 227.15741872787476}\n",
      "Losses {'ner': 244.19572138786316}\n",
      "Losses {'ner': 220.03549814224243}\n",
      "Losses {'ner': 225.72631454467773}\n",
      "Losses {'ner': 237.54323542118073}\n",
      "Losses {'ner': 222.27638912200928}\n",
      "Losses {'ner': 232.5384795665741}\n",
      "Losses {'ner': 228.8516867160797}\n",
      "Losses {'ner': 237.6918934583664}\n",
      "Losses {'ner': 229.97593212127686}\n",
      "Losses {'ner': 222.66355511546135}\n",
      "Losses {'ner': 232.9045855998993}\n",
      "Losses {'ner': 243.9825098514557}\n",
      "Losses {'ner': 222.42083060741425}\n",
      "Losses {'ner': 232.80055785179138}\n",
      "Losses {'ner': 228.9486496001482}\n",
      "Losses {'ner': 231.82639169692993}\n",
      "Losses {'ner': 220.65814125537872}\n",
      "Losses {'ner': 229.31019353866577}\n",
      "Losses {'ner': 224.9302191734314}\n",
      "Losses {'ner': 216.91144728660583}\n",
      "Losses {'ner': 225.16343450546265}\n",
      "Losses {'ner': 218.95722424983978}\n",
      "Losses {'ner': 220.16896012425423}\n",
      "Losses {'ner': 222.8226844072342}\n",
      "Losses {'ner': 240.28796291351318}\n",
      "Losses {'ner': 228.3137972354889}\n",
      "Losses {'ner': 230.45622849464417}\n",
      "Losses {'ner': 229.03823161125183}\n",
      "Losses {'ner': 225.44106245040894}\n",
      "Losses {'ner': 219.9954812526703}\n",
      "Losses {'ner': 226.65431118011475}\n",
      "Losses {'ner': 218.20777225494385}\n",
      "Losses {'ner': 230.93470633029938}\n",
      "Losses {'ner': 221.82668614387512}\n",
      "Losses {'ner': 241.722638130188}\n",
      "Losses {'ner': 229.45911812782288}\n",
      "Losses {'ner': 236.62870728969574}\n",
      "Losses {'ner': 220.9761998653412}\n",
      "Losses {'ner': 230.92732191085815}\n",
      "Losses {'ner': 220.84554052352905}\n",
      "Losses {'ner': 220.61639523506165}\n",
      "Losses {'ner': 230.821750998497}\n",
      "Losses {'ner': 226.28972864151}\n",
      "Losses {'ner': 235.2740604877472}\n",
      "Losses {'ner': 234.86836886405945}\n",
      "Losses {'ner': 229.13572144508362}\n",
      "Losses {'ner': 230.42826652526855}\n",
      "Losses {'ner': 231.32105408236384}\n",
      "Losses {'ner': 235.23746800422668}\n",
      "Losses {'ner': 224.44163942337036}\n",
      "Losses {'ner': 221.0948085784912}\n",
      "Losses {'ner': 223.90594577789307}\n",
      "Losses {'ner': 238.4893330335617}\n",
      "Losses {'ner': 227.3539559841156}\n",
      "Losses {'ner': 230.9338035583496}\n",
      "Losses {'ner': 227.78078854084015}\n",
      "Losses {'ner': 225.8368810415268}\n",
      "Losses {'ner': 232.73131775856018}\n",
      "Losses {'ner': 227.27975130081177}\n",
      "Losses {'ner': 224.87671899795532}\n",
      "Losses {'ner': 232.09940314292908}\n",
      "Losses {'ner': 225.0237033367157}\n",
      "Losses {'ner': 213.49602103233337}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 234.06029605865479}\n",
      "Losses {'ner': 224.44623839855194}\n",
      "Losses {'ner': 226.61991262435913}\n",
      "Losses {'ner': 228.2324914932251}\n",
      "Losses {'ner': 225.8400056362152}\n",
      "Losses {'ner': 215.78710782527924}\n",
      "Losses {'ner': 227.45232105255127}\n",
      "Losses {'ner': 232.13732707500458}\n",
      "Losses {'ner': 213.46120178699493}\n",
      "Losses {'ner': 224.65883660316467}\n",
      "Losses {'ner': 218.61355340480804}\n",
      "Losses {'ner': 228.58953285217285}\n",
      "Losses {'ner': 229.7282955646515}\n",
      "Losses {'ner': 230.84484493732452}\n",
      "Losses {'ner': 218.2988702058792}\n",
      "Losses {'ner': 228.2366235256195}\n",
      "Losses {'ner': 218.52835178375244}\n",
      "Losses {'ner': 223.44188010692596}\n",
      "Losses {'ner': 224.42252945899963}\n",
      "Losses {'ner': 231.42303562164307}\n",
      "Losses {'ner': 224.5454185605049}\n",
      "Losses {'ner': 222.4962478876114}\n",
      "Losses {'ner': 232.88351225852966}\n",
      "Losses {'ner': 223.95205986499786}\n",
      "Losses {'ner': 215.65498638153076}\n",
      "Losses {'ner': 218.73485386371613}\n",
      "Losses {'ner': 220.10327351093292}\n",
      "Losses {'ner': 225.18238615989685}\n",
      "Losses {'ner': 221.2974305152893}\n",
      "Losses {'ner': 220.65109610557556}\n",
      "Losses {'ner': 225.823632478714}\n",
      "Losses {'ner': 215.30198729038239}\n",
      "Losses {'ner': 221.182874917984}\n",
      "Losses {'ner': 231.95302379131317}\n",
      "Losses {'ner': 228.79077243804932}\n",
      "Losses {'ner': 216.60245406627655}\n",
      "Losses {'ner': 224.56338024139404}\n",
      "Losses {'ner': 225.47646176815033}\n",
      "Losses {'ner': 219.0920170545578}\n",
      "Losses {'ner': 223.98224294185638}\n",
      "Losses {'ner': 218.37384897470474}\n",
      "Losses {'ner': 234.1214997768402}\n",
      "Losses {'ner': 227.48478400707245}\n",
      "Losses {'ner': 243.7698051929474}\n",
      "Losses {'ner': 221.7288339138031}\n",
      "Losses {'ner': 213.41391777992249}\n",
      "Losses {'ner': 236.2149043083191}\n",
      "Losses {'ner': 211.14128303527832}\n",
      "Losses {'ner': 222.54839622974396}\n",
      "Losses {'ner': 221.6239562034607}\n",
      "Losses {'ner': 223.73117923736572}\n",
      "Losses {'ner': 212.1203225851059}\n",
      "Losses {'ner': 224.2116665840149}\n",
      "Losses {'ner': 219.36814785003662}\n",
      "Losses {'ner': 211.99476671218872}\n",
      "Losses {'ner': 234.51195192337036}\n",
      "Losses {'ner': 220.79419076442719}\n",
      "Losses {'ner': 228.24349164962769}\n",
      "Losses {'ner': 201.94029664993286}\n",
      "Losses {'ner': 215.67001390457153}\n",
      "Losses {'ner': 222.152965426445}\n",
      "Losses {'ner': 222.93119394779205}\n",
      "Losses {'ner': 213.39238321781158}\n",
      "Losses {'ner': 233.18071568012238}\n",
      "Losses {'ner': 223.0290343761444}\n",
      "Losses {'ner': 228.7653329372406}\n",
      "Losses {'ner': 215.36021733283997}\n",
      "Losses {'ner': 234.4249348640442}\n",
      "Losses {'ner': 236.30914950370789}\n",
      "Losses {'ner': 214.485582113266}\n",
      "Losses {'ner': 227.4414885044098}\n",
      "Losses {'ner': 214.15071058273315}\n",
      "Losses {'ner': 235.99010586738586}\n",
      "Losses {'ner': 228.44782829284668}\n",
      "Losses {'ner': 229.01968550682068}\n",
      "Losses {'ner': 201.33986645936966}\n",
      "Losses {'ner': 223.75253343582153}\n",
      "Losses {'ner': 224.07984817028046}\n",
      "Losses {'ner': 226.2235655784607}\n",
      "Losses {'ner': 207.91763353347778}\n",
      "Losses {'ner': 222.5056575536728}\n",
      "Losses {'ner': 214.6385973095894}\n",
      "Losses {'ner': 225.11031711101532}\n",
      "Losses {'ner': 215.20796793699265}\n",
      "Losses {'ner': 236.7504016160965}\n",
      "Losses {'ner': 220.53588318824768}\n",
      "Losses {'ner': 224.18106174468994}\n",
      "Losses {'ner': 236.631178855896}\n",
      "Losses {'ner': 221.11613327264786}\n",
      "Losses {'ner': 218.17958736419678}\n",
      "Losses {'ner': 226.41305100917816}\n",
      "Losses {'ner': 213.96946549415588}\n",
      "Losses {'ner': 226.35150122642517}\n",
      "Losses {'ner': 218.49709343910217}\n",
      "Losses {'ner': 235.72225564718246}\n",
      "Losses {'ner': 220.46732687950134}\n",
      "Losses {'ner': 226.9047611951828}\n",
      "Losses {'ner': 244.08290123939514}\n",
      "Losses {'ner': 228.38596785068512}\n",
      "Losses {'ner': 231.56269109249115}\n",
      "Losses {'ner': 210.37192803621292}\n",
      "Losses {'ner': 213.8104453086853}\n",
      "Losses {'ner': 228.85235059261322}\n",
      "Losses {'ner': 227.30724602937698}\n",
      "Losses {'ner': 221.9644067287445}\n",
      "Losses {'ner': 228.5116857290268}\n",
      "Losses {'ner': 231.58773469924927}\n",
      "Losses {'ner': 226.90210390090942}\n",
      "Losses {'ner': 218.4613072872162}\n",
      "Losses {'ner': 221.4762181043625}\n",
      "Losses {'ner': 222.42139518260956}\n",
      "Losses {'ner': 229.98839843273163}\n",
      "Losses {'ner': 226.34226608276367}\n",
      "Losses {'ner': 211.3365216255188}\n",
      "Losses {'ner': 207.4212462902069}\n",
      "Losses {'ner': 234.85241770744324}\n",
      "Losses {'ner': 218.61694598197937}\n",
      "Losses {'ner': 240.38939952850342}\n",
      "Losses {'ner': 210.2453976869583}\n",
      "Losses {'ner': 220.3947595357895}\n",
      "Losses {'ner': 221.20088720321655}\n",
      "Losses {'ner': 211.30504822731018}\n",
      "Losses {'ner': 227.83841609954834}\n",
      "Losses {'ner': 232.3631134033203}\n",
      "Losses {'ner': 232.22436720132828}\n",
      "Losses {'ner': 216.84308898448944}\n",
      "Losses {'ner': 225.965252161026}\n",
      "Losses {'ner': 217.9686959385872}\n",
      "Losses {'ner': 223.9784425497055}\n",
      "Losses {'ner': 227.64376187324524}\n",
      "Losses {'ner': 221.05448472499847}\n",
      "Losses {'ner': 222.9928914308548}\n",
      "Losses {'ner': 232.88765621185303}\n",
      "Losses {'ner': 216.51174473762512}\n",
      "Losses {'ner': 231.97549319267273}\n",
      "Losses {'ner': 231.2619252204895}\n",
      "Losses {'ner': 233.42228519916534}\n",
      "Losses {'ner': 223.31072306632996}\n",
      "Losses {'ner': 228.75140869617462}\n",
      "Losses {'ner': 222.10713386535645}\n",
      "Losses {'ner': 234.78154397010803}\n",
      "Losses {'ner': 221.45691341161728}\n",
      "Losses {'ner': 221.45520067214966}\n",
      "Losses {'ner': 227.2057830095291}\n",
      "Losses {'ner': 223.2163951396942}\n",
      "Losses {'ner': 212.16943216323853}\n",
      "Losses {'ner': 224.06636714935303}\n",
      "Losses {'ner': 218.51312470436096}\n",
      "Losses {'ner': 226.70656609535217}\n",
      "Losses {'ner': 224.8072212934494}\n",
      "Losses {'ner': 226.89422225952148}\n",
      "Losses {'ner': 231.2233282327652}\n",
      "Losses {'ner': 225.9987325668335}\n",
      "Losses {'ner': 244.40550756454468}\n",
      "Losses {'ner': 225.24948865175247}\n",
      "Losses {'ner': 233.7481106519699}\n",
      "Losses {'ner': 226.61879754066467}\n",
      "Losses {'ner': 218.2102769613266}\n",
      "Losses {'ner': 232.4152476787567}\n",
      "Losses {'ner': 225.02891874313354}\n",
      "Losses {'ner': 229.08640933036804}\n",
      "Losses {'ner': 215.58934020996094}\n",
      "Losses {'ner': 228.12248528003693}\n",
      "Losses {'ner': 224.67156398296356}\n",
      "Losses {'ner': 231.63981249928474}\n",
      "Losses {'ner': 226.0669651031494}\n",
      "Losses {'ner': 225.2789031267166}\n",
      "Losses {'ner': 231.27239155769348}\n",
      "Losses {'ner': 221.3978877067566}\n",
      "Losses {'ner': 235.16694688796997}\n",
      "Losses {'ner': 219.7201156616211}\n",
      "Losses {'ner': 220.83522629737854}\n",
      "Losses {'ner': 226.43439078330994}\n",
      "Losses {'ner': 233.91358923912048}\n",
      "Losses {'ner': 225.59841167926788}\n",
      "Losses {'ner': 218.97473239898682}\n",
      "Losses {'ner': 215.95227575302124}\n",
      "Losses {'ner': 224.68451976776123}\n",
      "Losses {'ner': 227.63068309426308}\n",
      "Losses {'ner': 224.88421726226807}\n",
      "Losses {'ner': 225.68667006492615}\n",
      "Losses {'ner': 211.94625174999237}\n",
      "Losses {'ner': 227.53408825397491}\n",
      "Losses {'ner': 211.12141346931458}\n",
      "Losses {'ner': 219.65066003799438}\n",
      "Losses {'ner': 237.2957513332367}\n",
      "Losses {'ner': 228.7884966135025}\n",
      "Losses {'ner': 234.21823596954346}\n",
      "Losses {'ner': 234.3908293247223}\n",
      "Losses {'ner': 216.8653456568718}\n",
      "Losses {'ner': 224.07181465625763}\n",
      "Losses {'ner': 235.8455195426941}\n",
      "Losses {'ner': 221.55282306671143}\n",
      "Losses {'ner': 209.87512755393982}\n",
      "Losses {'ner': 207.14089846611023}\n",
      "Losses {'ner': 230.6432341337204}\n",
      "Losses {'ner': 228.74976682662964}\n",
      "Losses {'ner': 230.03297954797745}\n",
      "Losses {'ner': 230.65111827850342}\n",
      "Losses {'ner': 228.50343024730682}\n",
      "Losses {'ner': 223.64932823181152}\n",
      "Losses {'ner': 218.2006597518921}\n",
      "Losses {'ner': 213.53500175476074}\n",
      "Losses {'ner': 218.72767162322998}\n",
      "Losses {'ner': 213.79165625572205}\n",
      "Losses {'ner': 221.14580941200256}\n",
      "Losses {'ner': 223.37595796585083}\n",
      "Losses {'ner': 228.02722215652466}\n",
      "Losses {'ner': 229.62139189243317}\n",
      "Losses {'ner': 217.79234313964844}\n",
      "Losses {'ner': 204.88632768392563}\n",
      "Losses {'ner': 232.81145703792572}\n",
      "Losses {'ner': 224.78448927402496}\n",
      "Losses {'ner': 218.5488988161087}\n",
      "Losses {'ner': 222.80498564243317}\n",
      "Losses {'ner': 234.077721118927}\n",
      "Losses {'ner': 224.21310877799988}\n",
      "Losses {'ner': 229.63279926776886}\n",
      "Losses {'ner': 223.46334755420685}\n",
      "Losses {'ner': 231.27684926986694}\n",
      "Losses {'ner': 224.23428785800934}\n",
      "Losses {'ner': 211.76519763469696}\n",
      "Losses {'ner': 233.7093870639801}\n",
      "Losses {'ner': 225.67713451385498}\n",
      "Losses {'ner': 224.63324666023254}\n",
      "Losses {'ner': 225.50675201416016}\n",
      "Losses {'ner': 213.30933499336243}\n",
      "Losses {'ner': 230.80687737464905}\n",
      "Losses {'ner': 231.1047124862671}\n",
      "Losses {'ner': 230.35821497440338}\n",
      "Losses {'ner': 224.64254474639893}\n",
      "Losses {'ner': 226.98053359985352}\n",
      "Losses {'ner': 233.70908427238464}\n",
      "Losses {'ner': 208.58719205856323}\n",
      "Losses {'ner': 231.35007286071777}\n",
      "Losses {'ner': 221.9988352060318}\n",
      "Losses {'ner': 221.04941606521606}\n",
      "Losses {'ner': 216.2900049686432}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 225.30700504779816}\n",
      "Losses {'ner': 217.28965175151825}\n",
      "Losses {'ner': 213.17387580871582}\n",
      "Losses {'ner': 224.89565467834473}\n",
      "Losses {'ner': 236.20264744758606}\n",
      "Losses {'ner': 224.15686058998108}\n",
      "Losses {'ner': 221.94922411441803}\n",
      "Losses {'ner': 234.90122663974762}\n",
      "Losses {'ner': 220.92245817184448}\n",
      "Losses {'ner': 227.03551524877548}\n",
      "Losses {'ner': 227.24245190620422}\n",
      "Losses {'ner': 221.51978504657745}\n",
      "Losses {'ner': 226.69247901439667}\n",
      "Losses {'ner': 237.3399441242218}\n",
      "Losses {'ner': 217.19330835342407}\n",
      "Losses {'ner': 230.26303911209106}\n",
      "Losses {'ner': 229.92630743980408}\n",
      "Losses {'ner': 228.28388690948486}\n",
      "Losses {'ner': 222.96378409862518}\n",
      "Losses {'ner': 215.8452788591385}\n",
      "Losses {'ner': 220.75283360481262}\n",
      "Losses {'ner': 214.617014169693}\n",
      "Losses {'ner': 210.1936045885086}\n",
      "Losses {'ner': 227.11062771081924}\n",
      "Losses {'ner': 228.4418615102768}\n",
      "Losses {'ner': 232.10828530788422}\n",
      "Losses {'ner': 222.00373470783234}\n",
      "Losses {'ner': 215.40457475185394}\n",
      "Losses {'ner': 220.07669913768768}\n",
      "Losses {'ner': 232.62982058525085}\n",
      "Losses {'ner': 225.17288041114807}\n",
      "Losses {'ner': 222.25529158115387}\n",
      "Losses {'ner': 220.19939124584198}\n",
      "Losses {'ner': 220.5426253080368}\n",
      "Losses {'ner': 223.59490764141083}\n",
      "Losses {'ner': 230.34785556793213}\n",
      "Losses {'ner': 228.21859407424927}\n",
      "Losses {'ner': 230.61980390548706}\n",
      "Losses {'ner': 224.9951912164688}\n",
      "Losses {'ner': 228.4486573934555}\n",
      "Losses {'ner': 217.3984715938568}\n",
      "Losses {'ner': 218.9506905078888}\n",
      "Losses {'ner': 229.10690808296204}\n",
      "Losses {'ner': 228.42984127998352}\n",
      "Losses {'ner': 228.9643131494522}\n",
      "Losses {'ner': 218.26807141304016}\n",
      "Losses {'ner': 221.25511980056763}\n",
      "Losses {'ner': 219.9186817407608}\n",
      "Losses {'ner': 239.60452544689178}\n",
      "Losses {'ner': 238.8536103963852}\n",
      "Losses {'ner': 212.5111949443817}\n",
      "Losses {'ner': 239.23157501220703}\n",
      "Losses {'ner': 226.50632965564728}\n",
      "Losses {'ner': 226.67929434776306}\n",
      "Losses {'ner': 230.5425522327423}\n",
      "Losses {'ner': 214.87314176559448}\n",
      "Losses {'ner': 218.73147106170654}\n",
      "Losses {'ner': 231.67472541332245}\n",
      "Losses {'ner': 218.25030493736267}\n",
      "Losses {'ner': 231.10214591026306}\n",
      "Losses {'ner': 221.12249994277954}\n",
      "Losses {'ner': 214.32691538333893}\n",
      "Losses {'ner': 216.57721948623657}\n",
      "Losses {'ner': 210.15127569437027}\n",
      "Losses {'ner': 216.12339067459106}\n",
      "Losses {'ner': 213.3474805355072}\n",
      "Losses {'ner': 229.54212856292725}\n",
      "Losses {'ner': 225.38775157928467}\n",
      "Losses {'ner': 213.83327615261078}\n",
      "Losses {'ner': 224.1250923871994}\n",
      "Losses {'ner': 234.7811747789383}\n",
      "Losses {'ner': 223.32406783103943}\n",
      "Losses {'ner': 224.77889919281006}\n",
      "Losses {'ner': 235.88183641433716}\n",
      "Losses {'ner': 229.2933225631714}\n",
      "Losses {'ner': 229.23824536800385}\n",
      "Losses {'ner': 212.96511198580265}\n",
      "Losses {'ner': 216.933249771595}\n",
      "Losses {'ner': 228.36171543598175}\n",
      "Losses {'ner': 215.58540749549866}\n",
      "Losses {'ner': 217.97522366046906}\n",
      "Losses {'ner': 215.06588411331177}\n",
      "Losses {'ner': 231.2873488664627}\n",
      "Losses {'ner': 225.19201385974884}\n",
      "Losses {'ner': 213.17508602142334}\n",
      "Losses {'ner': 223.44261705875397}\n",
      "Losses {'ner': 216.38139659166336}\n",
      "Losses {'ner': 204.75576949119568}\n",
      "Losses {'ner': 226.2106237411499}\n",
      "Losses {'ner': 226.37620568275452}\n",
      "Losses {'ner': 222.2669507265091}\n",
      "Losses {'ner': 227.19401490688324}\n",
      "Losses {'ner': 231.81249022483826}\n",
      "Losses {'ner': 227.84466862678528}\n",
      "Losses {'ner': 213.1259126663208}\n",
      "Losses {'ner': 234.76396596431732}\n",
      "Losses {'ner': 224.13947761058807}\n",
      "Losses {'ner': 222.3587999343872}\n",
      "Losses {'ner': 223.28919652104378}\n",
      "Losses {'ner': 245.0427701473236}\n",
      "Losses {'ner': 221.80854779481888}\n",
      "Losses {'ner': 222.95290648937225}\n",
      "Losses {'ner': 222.02027082443237}\n",
      "Losses {'ner': 226.6896369457245}\n",
      "Losses {'ner': 225.07541477680206}\n",
      "Losses {'ner': 227.58837151527405}\n",
      "Losses {'ner': 217.1314252614975}\n",
      "Losses {'ner': 225.8319971561432}\n",
      "Losses {'ner': 205.7374656200409}\n",
      "Losses {'ner': 230.92343664169312}\n",
      "Losses {'ner': 244.14689874649048}\n",
      "Losses {'ner': 232.16704976558685}\n",
      "Losses {'ner': 223.5032877922058}\n",
      "Losses {'ner': 232.2041209936142}\n",
      "Losses {'ner': 223.42134714126587}\n",
      "Losses {'ner': 229.35656905174255}\n",
      "Losses {'ner': 230.97565388679504}\n",
      "Losses {'ner': 225.71283316612244}\n",
      "Losses {'ner': 217.93863534927368}\n",
      "Losses {'ner': 222.67196643352509}\n",
      "Losses {'ner': 221.5797951221466}\n",
      "Losses {'ner': 219.21846163272858}\n",
      "Losses {'ner': 220.64198231697083}\n",
      "Losses {'ner': 225.40602147579193}\n",
      "Losses {'ner': 238.19756412506104}\n",
      "Losses {'ner': 239.78252625465393}\n",
      "Losses {'ner': 213.29882162809372}\n",
      "Losses {'ner': 226.70745885372162}\n",
      "Losses {'ner': 221.7616503238678}\n",
      "Losses {'ner': 221.65336346626282}\n",
      "Losses {'ner': 232.4784392118454}\n",
      "Losses {'ner': 228.78444623947144}\n",
      "Losses {'ner': 221.77476143836975}\n",
      "Losses {'ner': 228.67506980895996}\n",
      "Losses {'ner': 225.1675775051117}\n",
      "Losses {'ner': 225.3236637711525}\n",
      "Losses {'ner': 224.1234713792801}\n",
      "Losses {'ner': 225.43409824371338}\n",
      "Losses {'ner': 237.86229062080383}\n",
      "Losses {'ner': 230.5614082813263}\n",
      "Losses {'ner': 216.04476177692413}\n",
      "Losses {'ner': 232.36741185188293}\n",
      "Losses {'ner': 225.51301431655884}\n",
      "Losses {'ner': 224.0019313097}\n",
      "Losses {'ner': 233.07456922531128}\n",
      "Losses {'ner': 229.07350277900696}\n",
      "Losses {'ner': 232.595148563385}\n",
      "Losses {'ner': 225.2916260957718}\n",
      "Losses {'ner': 214.20982402563095}\n",
      "Losses {'ner': 228.54339182376862}\n",
      "Losses {'ner': 206.7744922041893}\n",
      "Losses {'ner': 221.56474995613098}\n",
      "Losses {'ner': 227.15131032466888}\n",
      "Losses {'ner': 217.4223232269287}\n",
      "Losses {'ner': 217.4390754699707}\n",
      "Losses {'ner': 230.8633826971054}\n",
      "Losses {'ner': 215.3543665409088}\n",
      "Losses {'ner': 232.26951384544373}\n",
      "Losses {'ner': 217.9580807685852}\n",
      "Losses {'ner': 227.01832175254822}\n",
      "Losses {'ner': 216.4682980775833}\n",
      "Losses {'ner': 229.06936836242676}\n",
      "Losses {'ner': 221.40183401107788}\n",
      "Losses {'ner': 219.46904075145721}\n",
      "Losses {'ner': 217.94024670124054}\n",
      "Losses {'ner': 227.57535004615784}\n",
      "Losses {'ner': 225.71070930361748}\n",
      "Losses {'ner': 228.03275394439697}\n",
      "Losses {'ner': 219.27616620063782}\n",
      "Losses {'ner': 218.80424296855927}\n",
      "Losses {'ner': 234.43930423259735}\n",
      "Losses {'ner': 226.4721338748932}\n",
      "Losses {'ner': 234.63454747200012}\n",
      "Losses {'ner': 218.70071506500244}\n",
      "Losses {'ner': 232.47278916835785}\n",
      "Losses {'ner': 220.7883095741272}\n",
      "Losses {'ner': 219.21845698356628}\n",
      "Losses {'ner': 227.13622188568115}\n",
      "Losses {'ner': 221.54145348072052}\n",
      "Losses {'ner': 220.3579819202423}\n",
      "Losses {'ner': 210.93722969293594}\n",
      "Losses {'ner': 204.93974614143372}\n",
      "Losses {'ner': 229.69042921066284}\n",
      "Losses {'ner': 233.1088343858719}\n",
      "Losses {'ner': 230.88756585121155}\n",
      "Losses {'ner': 227.54903066158295}\n",
      "Losses {'ner': 220.1258932352066}\n",
      "Losses {'ner': 211.49915766716003}\n",
      "Losses {'ner': 225.63867151737213}\n",
      "Losses {'ner': 218.99702835083008}\n",
      "Losses {'ner': 227.91616988182068}\n",
      "Losses {'ner': 232.50224101543427}\n",
      "Losses {'ner': 228.19410502910614}\n",
      "Losses {'ner': 207.60791981220245}\n",
      "Losses {'ner': 222.33166432380676}\n",
      "Losses {'ner': 230.8631809949875}\n",
      "Losses {'ner': 235.7746343612671}\n",
      "Losses {'ner': 226.48045361042023}\n",
      "Losses {'ner': 234.46671438217163}\n",
      "Losses {'ner': 215.124467253685}\n",
      "Losses {'ner': 213.96667385101318}\n",
      "Losses {'ner': 228.06372570991516}\n",
      "Losses {'ner': 231.66069054603577}\n",
      "Losses {'ner': 216.713969707489}\n",
      "Losses {'ner': 226.88640761375427}\n",
      "Losses {'ner': 212.08853495121002}\n",
      "Losses {'ner': 229.70180529356003}\n",
      "Losses {'ner': 234.7215178012848}\n",
      "Losses {'ner': 234.3107533454895}\n",
      "Losses {'ner': 233.47767734527588}\n",
      "Losses {'ner': 216.30062079429626}\n",
      "Losses {'ner': 229.65210103988647}\n",
      "Losses {'ner': 228.43491160869598}\n",
      "Losses {'ner': 226.25503778457642}\n",
      "Losses {'ner': 215.5856120288372}\n",
      "Losses {'ner': 233.11636424064636}\n",
      "Losses {'ner': 220.4767460823059}\n",
      "Losses {'ner': 222.3681559562683}\n",
      "Losses {'ner': 218.48940634727478}\n",
      "Losses {'ner': 218.09915852546692}\n",
      "Losses {'ner': 221.66820377111435}\n",
      "Losses {'ner': 216.67077541351318}\n",
      "Losses {'ner': 232.41038632392883}\n",
      "Losses {'ner': 209.82553839683533}\n",
      "Losses {'ner': 214.40077137947083}\n",
      "Losses {'ner': 219.21627604961395}\n",
      "Losses {'ner': 225.3167808651924}\n",
      "Losses {'ner': 221.7746146917343}\n",
      "Losses {'ner': 219.41869151592255}\n",
      "Losses {'ner': 218.48171949386597}\n",
      "Losses {'ner': 214.11456489562988}\n",
      "Losses {'ner': 222.77898514270782}\n",
      "Losses {'ner': 220.452632188797}\n",
      "Losses {'ner': 224.00985479354858}\n",
      "Losses {'ner': 219.73181986808777}\n",
      "Losses {'ner': 222.1503962278366}\n",
      "Losses {'ner': 234.0354516506195}\n",
      "Losses {'ner': 216.63265240192413}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 213.61089837551117}\n",
      "Losses {'ner': 228.1989507675171}\n",
      "Losses {'ner': 222.84538543224335}\n",
      "Losses {'ner': 222.60237288475037}\n",
      "Losses {'ner': 233.0583107471466}\n",
      "Losses {'ner': 223.16765892505646}\n",
      "Losses {'ner': 220.6858479976654}\n",
      "Losses {'ner': 219.16819179058075}\n",
      "Losses {'ner': 219.21430850028992}\n",
      "Losses {'ner': 238.12966322898865}\n",
      "Losses {'ner': 232.30180323123932}\n",
      "Losses {'ner': 218.96061658859253}\n",
      "Losses {'ner': 228.3099434375763}\n",
      "Losses {'ner': 211.9642072916031}\n",
      "Losses {'ner': 225.54283320903778}\n",
      "Losses {'ner': 235.92019057273865}\n",
      "Losses {'ner': 211.53769731521606}\n",
      "Losses {'ner': 223.74415040016174}\n",
      "Losses {'ner': 220.81981706619263}\n",
      "Losses {'ner': 228.04955196380615}\n",
      "Losses {'ner': 227.37063241004944}\n",
      "Losses {'ner': 199.87327069044113}\n",
      "Losses {'ner': 231.41696667671204}\n",
      "Losses {'ner': 234.31311106681824}\n",
      "Losses {'ner': 225.17151254415512}\n",
      "Losses {'ner': 230.1159611940384}\n",
      "Losses {'ner': 234.63792753219604}\n",
      "Losses {'ner': 225.21848249435425}\n",
      "Losses {'ner': 229.4640210866928}\n",
      "Losses {'ner': 227.45262122154236}\n",
      "Losses {'ner': 205.65083122253418}\n",
      "Losses {'ner': 220.26193416118622}\n",
      "Losses {'ner': 219.3010904788971}\n",
      "Losses {'ner': 228.7236728668213}\n",
      "Losses {'ner': 223.9351134300232}\n",
      "Losses {'ner': 235.5357096195221}\n",
      "Losses {'ner': 223.44910943508148}\n",
      "Losses {'ner': 213.5866494178772}\n",
      "Losses {'ner': 226.0335726737976}\n",
      "Losses {'ner': 210.8576943874359}\n",
      "Losses {'ner': 229.2012995481491}\n",
      "Losses {'ner': 218.44140243530273}\n",
      "Losses {'ner': 232.27177095413208}\n",
      "Losses {'ner': 214.7045133113861}\n",
      "Losses {'ner': 231.47253382205963}\n",
      "Losses {'ner': 213.96463051438332}\n",
      "Losses {'ner': 218.21522128582}\n",
      "Losses {'ner': 226.48675167560577}\n",
      "Losses {'ner': 236.45664167404175}\n",
      "Losses {'ner': 227.65461993217468}\n",
      "Losses {'ner': 227.66587162017822}\n",
      "Losses {'ner': 215.45904779434204}\n",
      "Losses {'ner': 217.36152744293213}\n",
      "Losses {'ner': 234.9711558818817}\n",
      "Losses {'ner': 222.6929473876953}\n",
      "Losses {'ner': 223.82348144054413}\n",
      "Losses {'ner': 241.01017904281616}\n",
      "Losses {'ner': 211.11903858184814}\n",
      "Losses {'ner': 216.74396389722824}\n",
      "Losses {'ner': 229.31193840503693}\n",
      "Losses {'ner': 212.34777587652206}\n",
      "Losses {'ner': 227.0691990852356}\n",
      "Losses {'ner': 224.64924597740173}\n",
      "Losses {'ner': 213.79618346691132}\n",
      "Losses {'ner': 228.63288068771362}\n",
      "Losses {'ner': 224.0269160270691}\n",
      "Losses {'ner': 226.78007292747498}\n",
      "Losses {'ner': 228.26367783546448}\n",
      "Losses {'ner': 213.0267276763916}\n",
      "Losses {'ner': 221.2148683667183}\n",
      "Losses {'ner': 214.7964278459549}\n",
      "Losses {'ner': 226.7726507782936}\n",
      "Losses {'ner': 213.46486377716064}\n",
      "Losses {'ner': 219.50515246391296}\n",
      "Losses {'ner': 233.0292980670929}\n",
      "Losses {'ner': 228.78259027004242}\n",
      "Losses {'ner': 221.18552899360657}\n",
      "Losses {'ner': 226.8815038204193}\n",
      "Losses {'ner': 229.50480163097382}\n",
      "Losses {'ner': 219.93690359592438}\n",
      "Losses {'ner': 229.17649203538895}\n",
      "Losses {'ner': 229.87752270698547}\n",
      "Losses {'ner': 219.3146606683731}\n",
      "Losses {'ner': 225.23214197158813}\n",
      "Losses {'ner': 222.22547435760498}\n",
      "Losses {'ner': 222.04445958137512}\n",
      "Losses {'ner': 216.5125311613083}\n",
      "Losses {'ner': 222.16231071949005}\n",
      "Losses {'ner': 234.51694297790527}\n",
      "Losses {'ner': 225.82383251190186}\n",
      "Losses {'ner': 225.32382768392563}\n",
      "Losses {'ner': 237.76925539970398}\n",
      "Losses {'ner': 230.30320620536804}\n",
      "Losses {'ner': 217.38351714611053}\n",
      "Losses {'ner': 219.14615444839}\n",
      "Losses {'ner': 221.42910408973694}\n",
      "Losses {'ner': 227.34564185142517}\n",
      "Losses {'ner': 223.2555763721466}\n",
      "Losses {'ner': 222.290301322937}\n",
      "Losses {'ner': 215.25838708877563}\n",
      "Losses {'ner': 208.71106815338135}\n",
      "Losses {'ner': 227.48980522155762}\n",
      "Losses {'ner': 218.12804552167654}\n",
      "Losses {'ner': 223.39904272556305}\n",
      "Losses {'ner': 226.0155348777771}\n",
      "Losses {'ner': 211.22870874404907}\n",
      "Losses {'ner': 226.6415069103241}\n",
      "Losses {'ner': 219.67709183692932}\n",
      "Losses {'ner': 225.90284287929535}\n",
      "Losses {'ner': 238.2965865135193}\n",
      "Losses {'ner': 234.81739377975464}\n",
      "Losses {'ner': 221.32194072008133}\n",
      "Losses {'ner': 225.3085880279541}\n",
      "Losses {'ner': 212.75416219234467}\n",
      "Losses {'ner': 218.5379958152771}\n",
      "Losses {'ner': 223.78337943553925}\n",
      "Losses {'ner': 235.5711648464203}\n",
      "Losses {'ner': 221.27774024009705}\n",
      "Losses {'ner': 224.47715854644775}\n",
      "Losses {'ner': 223.86724030971527}\n",
      "Losses {'ner': 227.09200620651245}\n",
      "Losses {'ner': 222.92778420448303}\n",
      "Losses {'ner': 224.69764471054077}\n",
      "Losses {'ner': 221.8645932674408}\n",
      "Losses {'ner': 222.12549686431885}\n",
      "Losses {'ner': 222.81448769569397}\n",
      "Losses {'ner': 210.2870762348175}\n",
      "Losses {'ner': 221.19672739505768}\n",
      "Losses {'ner': 228.86628341674805}\n",
      "Losses {'ner': 239.7163804769516}\n",
      "Losses {'ner': 224.57453322410583}\n",
      "Losses {'ner': 218.796492934227}\n",
      "Losses {'ner': 229.77791738510132}\n",
      "Losses {'ner': 232.5453929901123}\n",
      "Losses {'ner': 215.55541467666626}\n",
      "Losses {'ner': 230.30923736095428}\n",
      "Losses {'ner': 231.11182141304016}\n",
      "Losses {'ner': 226.6208188533783}\n",
      "Losses {'ner': 218.76627135276794}\n",
      "Losses {'ner': 221.75429439544678}\n",
      "Losses {'ner': 228.30789232254028}\n",
      "Losses {'ner': 215.24154901504517}\n",
      "Losses {'ner': 207.4050377011299}\n",
      "Losses {'ner': 221.63067018985748}\n",
      "Losses {'ner': 224.94219362735748}\n",
      "Losses {'ner': 216.111785531044}\n",
      "Losses {'ner': 221.44701218605042}\n",
      "Losses {'ner': 225.0640219449997}\n",
      "Losses {'ner': 228.22074913978577}\n",
      "Losses {'ner': 220.976881980896}\n",
      "Losses {'ner': 214.24995702505112}\n",
      "Losses {'ner': 218.39017629623413}\n",
      "Losses {'ner': 228.539874792099}\n",
      "Losses {'ner': 247.0194046497345}\n",
      "Losses {'ner': 220.25619840621948}\n",
      "Losses {'ner': 234.74022269248962}\n",
      "Losses {'ner': 234.8397512435913}\n",
      "Losses {'ner': 228.99067974090576}\n",
      "Losses {'ner': 213.41753792762756}\n",
      "Losses {'ner': 216.03099691867828}\n",
      "Losses {'ner': 230.0323166847229}\n",
      "Losses {'ner': 224.4256672859192}\n",
      "Losses {'ner': 224.23435163497925}\n",
      "Losses {'ner': 216.41225218772888}\n",
      "Losses {'ner': 229.63886284828186}\n",
      "Losses {'ner': 215.33204591274261}\n",
      "Losses {'ner': 216.84976065158844}\n",
      "Losses {'ner': 228.1489450931549}\n",
      "Losses {'ner': 239.22463178634644}\n",
      "Losses {'ner': 209.21340548992157}\n",
      "Losses {'ner': 220.46794283390045}\n",
      "Losses {'ner': 219.2221314907074}\n",
      "Losses {'ner': 231.61176896095276}\n",
      "Losses {'ner': 217.38944094628096}\n",
      "Losses {'ner': 215.67990255355835}\n",
      "Losses {'ner': 219.30010080337524}\n",
      "Losses {'ner': 206.08645844459534}\n",
      "Losses {'ner': 227.13311755657196}\n",
      "Losses {'ner': 219.10609829425812}\n",
      "Losses {'ner': 224.0105917453766}\n",
      "Losses {'ner': 210.7170684337616}\n",
      "Losses {'ner': 218.31176233291626}\n",
      "Losses {'ner': 214.21809840202332}\n",
      "Losses {'ner': 201.82549971342087}\n",
      "Losses {'ner': 228.24268117547035}\n",
      "Losses {'ner': 219.46599650382996}\n",
      "Losses {'ner': 211.30750644207}\n",
      "Losses {'ner': 230.07252991199493}\n",
      "Losses {'ner': 224.7935413122177}\n",
      "Losses {'ner': 226.48914003372192}\n",
      "Losses {'ner': 228.81258702278137}\n",
      "Losses {'ner': 225.944673538208}\n",
      "Losses {'ner': 223.02493834495544}\n",
      "Losses {'ner': 219.4470740556717}\n",
      "Losses {'ner': 220.61757862567902}\n",
      "Losses {'ner': 222.0252182483673}\n",
      "Losses {'ner': 228.46298027038574}\n",
      "Losses {'ner': 233.60138082504272}\n",
      "Losses {'ner': 230.55045148730278}\n",
      "Losses {'ner': 225.82502579689026}\n",
      "Losses {'ner': 219.9013409614563}\n",
      "Losses {'ner': 232.4070544242859}\n",
      "Losses {'ner': 233.96037983894348}\n",
      "Losses {'ner': 215.98859848082066}\n",
      "Losses {'ner': 226.33075404167175}\n",
      "Losses {'ner': 236.1230230331421}\n",
      "Losses {'ner': 211.12823963165283}\n",
      "Losses {'ner': 223.33598566055298}\n",
      "Losses {'ner': 227.83483147621155}\n",
      "Losses {'ner': 222.7243082523346}\n",
      "Losses {'ner': 208.28562545776367}\n",
      "Losses {'ner': 229.19564032554626}\n",
      "Losses {'ner': 227.31039987504482}\n",
      "Losses {'ner': 231.1864252090454}\n",
      "Losses {'ner': 228.47170758247375}\n",
      "Losses {'ner': 231.38531041145325}\n",
      "Losses {'ner': 229.19969773292542}\n",
      "Losses {'ner': 216.6070671081543}\n",
      "Losses {'ner': 212.15405535697937}\n",
      "Losses {'ner': 221.4562450647354}\n",
      "Losses {'ner': 213.51857447624207}\n",
      "Losses {'ner': 217.91047883033752}\n",
      "Losses {'ner': 219.17840325832367}\n",
      "Losses {'ner': 217.17552709579468}\n",
      "Losses {'ner': 220.01639795303345}\n",
      "Losses {'ner': 220.83113807439804}\n",
      "Losses {'ner': 215.72211050987244}\n",
      "Losses {'ner': 229.5135853290558}\n",
      "Losses {'ner': 222.83127105236053}\n",
      "Losses {'ner': 211.91221344470978}\n",
      "Losses {'ner': 210.12919354438782}\n",
      "Losses {'ner': 231.2758436203003}\n",
      "Losses {'ner': 224.0445081591606}\n",
      "Losses {'ner': 216.34944665431976}\n",
      "Losses {'ner': 218.2394106388092}\n",
      "Losses {'ner': 218.6870095729828}\n",
      "Losses {'ner': 223.5085484981537}\n",
      "Losses {'ner': 213.6746895313263}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 225.86309790611267}\n",
      "Losses {'ner': 239.23623794317245}\n",
      "Losses {'ner': 232.07323372364044}\n",
      "Losses {'ner': 223.09524369239807}\n",
      "Losses {'ner': 226.5610911846161}\n",
      "Losses {'ner': 230.08934330940247}\n",
      "Losses {'ner': 229.33655261993408}\n",
      "Losses {'ner': 219.05677264928818}\n",
      "Losses {'ner': 215.13153910636902}\n",
      "Losses {'ner': 229.40422415733337}\n",
      "Losses {'ner': 224.1802854537964}\n",
      "Losses {'ner': 233.11382293701172}\n",
      "Losses {'ner': 211.28584066033363}\n",
      "Losses {'ner': 229.0874834060669}\n",
      "Losses {'ner': 216.67899841070175}\n",
      "Losses {'ner': 219.18678283691406}\n",
      "Losses {'ner': 219.06195044517517}\n",
      "Losses {'ner': 224.19118475914001}\n",
      "Losses {'ner': 218.24024891853333}\n",
      "Losses {'ner': 228.57582640647888}\n",
      "Losses {'ner': 227.54847526550293}\n",
      "Losses {'ner': 228.74887466430664}\n",
      "Losses {'ner': 230.56358742713928}\n",
      "Losses {'ner': 210.02797836065292}\n",
      "Losses {'ner': 210.00681972503662}\n",
      "Losses {'ner': 225.59749746322632}\n",
      "Losses {'ner': 218.6946816444397}\n",
      "Losses {'ner': 243.63075351715088}\n",
      "Losses {'ner': 244.2168378829956}\n",
      "Losses {'ner': 240.3805296421051}\n",
      "Losses {'ner': 220.81063532829285}\n",
      "Losses {'ner': 225.14448142051697}\n",
      "Losses {'ner': 235.29220986366272}\n",
      "Losses {'ner': 226.87615704536438}\n",
      "Losses {'ner': 220.45313215255737}\n",
      "Losses {'ner': 231.04256939888}\n",
      "Losses {'ner': 226.92395639419556}\n",
      "Losses {'ner': 230.59108352661133}\n",
      "Losses {'ner': 224.5384566783905}\n",
      "Losses {'ner': 226.98200678825378}\n",
      "Losses {'ner': 217.4609706401825}\n",
      "Losses {'ner': 230.5776252746582}\n",
      "Losses {'ner': 217.11249697208405}\n",
      "Losses {'ner': 232.35113954544067}\n",
      "Losses {'ner': 225.14414310455322}\n",
      "Losses {'ner': 231.83009868860245}\n",
      "Losses {'ner': 229.61933636665344}\n",
      "Losses {'ner': 215.72878205776215}\n",
      "Losses {'ner': 228.22517585754395}\n",
      "Losses {'ner': 223.4098720550537}\n",
      "Losses {'ner': 227.12632369995117}\n",
      "Losses {'ner': 222.88876867294312}\n",
      "Losses {'ner': 224.0828014612198}\n",
      "Losses {'ner': 215.47556459903717}\n",
      "Losses {'ner': 219.4715874195099}\n",
      "Losses {'ner': 226.67064368724823}\n",
      "Losses {'ner': 217.89186787605286}\n",
      "Losses {'ner': 225.48656356334686}\n",
      "Losses {'ner': 229.81480145454407}\n",
      "Losses {'ner': 222.10112309455872}\n",
      "Losses {'ner': 202.53342020511627}\n",
      "Losses {'ner': 217.975688457489}\n",
      "Losses {'ner': 224.76558327674866}\n",
      "Losses {'ner': 213.01622545719147}\n",
      "Losses {'ner': 232.24155712127686}\n",
      "Losses {'ner': 217.2771600484848}\n",
      "Losses {'ner': 222.1172308921814}\n",
      "Losses {'ner': 234.66062378883362}\n",
      "Losses {'ner': 218.78072774410248}\n",
      "Losses {'ner': 206.50189411640167}\n",
      "Losses {'ner': 205.46306729316711}\n",
      "Losses {'ner': 229.49711346626282}\n",
      "Losses {'ner': 215.767143368721}\n",
      "Losses {'ner': 219.2373350262642}\n",
      "Losses {'ner': 236.48973035812378}\n",
      "Losses {'ner': 225.0682134628296}\n",
      "Losses {'ner': 228.1479994058609}\n",
      "Losses {'ner': 220.01009964942932}\n",
      "Losses {'ner': 199.76121163368225}\n",
      "Losses {'ner': 230.8866982460022}\n",
      "Losses {'ner': 218.8956642150879}\n",
      "Losses {'ner': 223.6021921634674}\n",
      "Losses {'ner': 203.97975552082062}\n",
      "Losses {'ner': 219.5119330883026}\n",
      "Losses {'ner': 218.7230987548828}\n",
      "Losses {'ner': 235.28080487251282}\n",
      "Losses {'ner': 217.19817423820496}\n",
      "Losses {'ner': 222.72895550727844}\n",
      "Losses {'ner': 220.53611302375793}\n",
      "Losses {'ner': 220.18972849845886}\n",
      "Losses {'ner': 220.43283224105835}\n",
      "Losses {'ner': 235.8049577474594}\n",
      "Losses {'ner': 223.39173305034637}\n",
      "Losses {'ner': 231.57309007644653}\n",
      "Losses {'ner': 221.74860954284668}\n",
      "Losses {'ner': 212.94069874286652}\n",
      "Losses {'ner': 220.90891122817993}\n",
      "Losses {'ner': 216.49126172065735}\n",
      "Losses {'ner': 221.6025893688202}\n",
      "Losses {'ner': 231.18481755256653}\n",
      "Losses {'ner': 223.84156227111816}\n",
      "Losses {'ner': 226.04151809215546}\n",
      "Losses {'ner': 214.76887667179108}\n",
      "Losses {'ner': 228.92188811302185}\n",
      "Losses {'ner': 218.14424419403076}\n",
      "Losses {'ner': 211.07514548301697}\n",
      "Losses {'ner': 217.20382606983185}\n",
      "Losses {'ner': 223.40547633171082}\n",
      "Losses {'ner': 217.81663060188293}\n",
      "Losses {'ner': 225.04491762816906}\n",
      "Losses {'ner': 224.4932797551155}\n",
      "Losses {'ner': 220.08610951900482}\n",
      "Losses {'ner': 225.3001276254654}\n",
      "Losses {'ner': 219.597154378891}\n",
      "Losses {'ner': 233.7755947113037}\n",
      "Losses {'ner': 214.19011402130127}\n",
      "Losses {'ner': 218.71804666519165}\n",
      "Losses {'ner': 219.04177808761597}\n",
      "Losses {'ner': 217.44878590106964}\n",
      "Losses {'ner': 213.74967169761658}\n",
      "Losses {'ner': 223.78740918636322}\n",
      "Losses {'ner': 226.67102146148682}\n",
      "Losses {'ner': 229.5755764245987}\n",
      "Losses {'ner': 218.58715760707855}\n",
      "Losses {'ner': 220.70775890350342}\n",
      "Losses {'ner': 228.96373224258423}\n",
      "Losses {'ner': 221.46520590782166}\n",
      "Losses {'ner': 226.40864551067352}\n",
      "Losses {'ner': 212.74940085411072}\n",
      "Losses {'ner': 211.42928206920624}\n",
      "Losses {'ner': 216.35874927043915}\n",
      "Losses {'ner': 216.22477638721466}\n",
      "Losses {'ner': 227.43245029449463}\n",
      "Losses {'ner': 222.25062429904938}\n",
      "Losses {'ner': 237.1642097234726}\n",
      "Losses {'ner': 220.249577999115}\n",
      "Losses {'ner': 211.6017097234726}\n",
      "Losses {'ner': 226.73265504837036}\n",
      "Losses {'ner': 235.30225610733032}\n",
      "Losses {'ner': 223.21148812770844}\n",
      "Losses {'ner': 207.52772068977356}\n",
      "Losses {'ner': 218.65761733055115}\n",
      "Losses {'ner': 215.45577335357666}\n",
      "Losses {'ner': 215.78110671043396}\n",
      "Losses {'ner': 210.76969504356384}\n",
      "Losses {'ner': 218.82828402519226}\n",
      "Losses {'ner': 222.79110980033875}\n",
      "Losses {'ner': 221.50147533416748}\n",
      "Losses {'ner': 216.72500717639923}\n",
      "Losses {'ner': 226.85771870613098}\n",
      "Losses {'ner': 217.87924885749817}\n",
      "Losses {'ner': 219.3684927225113}\n",
      "Losses {'ner': 223.44854760169983}\n",
      "Losses {'ner': 211.15368002653122}\n",
      "Losses {'ner': 229.4122918844223}\n",
      "Losses {'ner': 211.289302110672}\n",
      "Losses {'ner': 220.53453421592712}\n",
      "Losses {'ner': 225.51815247535706}\n",
      "Losses {'ner': 235.91632056236267}\n",
      "Losses {'ner': 232.29786729812622}\n",
      "Losses {'ner': 224.5589635372162}\n",
      "Losses {'ner': 228.87206614017487}\n",
      "Losses {'ner': 225.48606967926025}\n",
      "Losses {'ner': 221.60253608226776}\n",
      "Losses {'ner': 222.50349175930023}\n",
      "Losses {'ner': 215.3590931892395}\n",
      "Losses {'ner': 230.78195452690125}\n",
      "Losses {'ner': 228.09316980838776}\n",
      "Losses {'ner': 221.5692617893219}\n",
      "Losses {'ner': 222.41264808177948}\n",
      "Losses {'ner': 221.58028376102448}\n",
      "Losses {'ner': 227.450914144516}\n",
      "Losses {'ner': 229.8378357887268}\n",
      "Losses {'ner': 222.264599442482}\n",
      "Losses {'ner': 223.13427555561066}\n",
      "Losses {'ner': 224.4923998117447}\n",
      "Losses {'ner': 218.8262267112732}\n",
      "Losses {'ner': 232.77681624889374}\n",
      "Losses {'ner': 231.67310237884521}\n",
      "Losses {'ner': 235.99791979789734}\n",
      "Losses {'ner': 226.7848893404007}\n",
      "Losses {'ner': 220.7833687067032}\n",
      "Losses {'ner': 225.50323903560638}\n",
      "Losses {'ner': 221.23589271306992}\n",
      "Losses {'ner': 237.04921317100525}\n",
      "Losses {'ner': 220.00120770931244}\n",
      "Losses {'ner': 218.98379409313202}\n",
      "Losses {'ner': 236.7543957233429}\n",
      "Losses {'ner': 232.54812860488892}\n",
      "Losses {'ner': 226.04375410079956}\n",
      "Losses {'ner': 204.8801884651184}\n",
      "Losses {'ner': 237.97633695602417}\n",
      "Losses {'ner': 202.7122504711151}\n",
      "Losses {'ner': 232.66480159759521}\n",
      "Losses {'ner': 226.8716266155243}\n",
      "Losses {'ner': 222.23088896274567}\n",
      "Losses {'ner': 203.53415846824646}\n",
      "Losses {'ner': 221.9437299966812}\n",
      "Losses {'ner': 224.41314673423767}\n",
      "Losses {'ner': 212.96518528461456}\n",
      "Losses {'ner': 222.38987731933594}\n",
      "Losses {'ner': 209.49347364902496}\n",
      "Losses {'ner': 226.96442782878876}\n",
      "Losses {'ner': 233.83091342449188}\n",
      "Losses {'ner': 219.00664600729942}\n",
      "Losses {'ner': 214.42747700214386}\n",
      "Losses {'ner': 231.1986232995987}\n",
      "Losses {'ner': 207.58770608901978}\n",
      "Losses {'ner': 226.06884956359863}\n",
      "Losses {'ner': 217.29846823215485}\n",
      "Losses {'ner': 214.13873422145844}\n",
      "Losses {'ner': 223.30511116981506}\n",
      "Losses {'ner': 216.55347788333893}\n",
      "Losses {'ner': 228.80941343307495}\n",
      "Losses {'ner': 226.18447935581207}\n",
      "Losses {'ner': 227.72955399751663}\n",
      "Losses {'ner': 230.0834310054779}\n",
      "Losses {'ner': 228.5289969444275}\n",
      "Losses {'ner': 213.5753300189972}\n",
      "Losses {'ner': 221.85704278945923}\n",
      "Losses {'ner': 225.26252162456512}\n",
      "Losses {'ner': 227.48525726795197}\n",
      "Losses {'ner': 224.61629569530487}\n",
      "Losses {'ner': 221.74066352844238}\n",
      "Losses {'ner': 209.49267268180847}\n",
      "Losses {'ner': 222.1670345067978}\n",
      "Losses {'ner': 224.4495724439621}\n",
      "Losses {'ner': 219.3259493112564}\n",
      "Losses {'ner': 216.10060894489288}\n",
      "Losses {'ner': 216.45208191871643}\n",
      "Losses {'ner': 225.42126941680908}\n",
      "Losses {'ner': 220.95922935009003}\n",
      "Losses {'ner': 209.09371328353882}\n",
      "Losses {'ner': 230.74206793308258}\n",
      "Losses {'ner': 229.14523142576218}\n",
      "Losses {'ner': 227.81970596313477}\n",
      "Losses {'ner': 214.65479731559753}\n",
      "Losses {'ner': 237.28751611709595}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 227.89938914775848}\n",
      "Losses {'ner': 220.1625857949257}\n",
      "Losses {'ner': 208.21529364585876}\n",
      "Losses {'ner': 230.65483993291855}\n",
      "Losses {'ner': 228.53335082530975}\n",
      "Losses {'ner': 216.69419407844543}\n",
      "Losses {'ner': 233.9569103717804}\n",
      "Losses {'ner': 220.39820313453674}\n",
      "Losses {'ner': 216.38091373443604}\n",
      "Losses {'ner': 229.4251767396927}\n",
      "Losses {'ner': 216.0474190711975}\n",
      "Losses {'ner': 222.68388271331787}\n",
      "Losses {'ner': 214.70230412483215}\n",
      "Losses {'ner': 235.71457666158676}\n",
      "Losses {'ner': 215.2385675907135}\n",
      "Losses {'ner': 220.2413592338562}\n",
      "Losses {'ner': 210.9160588979721}\n",
      "Losses {'ner': 215.9551488161087}\n",
      "Losses {'ner': 227.68841588497162}\n",
      "Losses {'ner': 228.97592389583588}\n",
      "Losses {'ner': 206.1538963317871}\n",
      "Losses {'ner': 205.78163766860962}\n",
      "Losses {'ner': 217.77827787399292}\n",
      "Losses {'ner': 227.64248204231262}\n",
      "Losses {'ner': 218.47804927825928}\n",
      "Losses {'ner': 219.7745600938797}\n",
      "Losses {'ner': 215.26652002334595}\n",
      "Losses {'ner': 225.3178265094757}\n",
      "Losses {'ner': 220.85143527388573}\n",
      "Losses {'ner': 233.55127930641174}\n",
      "Losses {'ner': 214.18864834308624}\n",
      "Losses {'ner': 227.87135910987854}\n",
      "Losses {'ner': 234.30466866493225}\n",
      "Losses {'ner': 225.3830189704895}\n",
      "Losses {'ner': 241.18881607055664}\n",
      "Losses {'ner': 210.49198424816132}\n",
      "Losses {'ner': 209.89965963363647}\n",
      "Losses {'ner': 217.84417712688446}\n",
      "Losses {'ner': 236.1743869781494}\n",
      "Losses {'ner': 228.8346619606018}\n",
      "Losses {'ner': 223.0006264448166}\n",
      "Losses {'ner': 220.02792501449585}\n",
      "Losses {'ner': 222.53836458921432}\n",
      "Losses {'ner': 217.83749437332153}\n",
      "Losses {'ner': 230.43813920021057}\n",
      "Losses {'ner': 217.9111659526825}\n",
      "Losses {'ner': 224.46754145622253}\n",
      "Losses {'ner': 222.74994659423828}\n",
      "Losses {'ner': 232.95861399173737}\n",
      "Losses {'ner': 202.1287466287613}\n",
      "Losses {'ner': 230.33328247070312}\n",
      "Losses {'ner': 230.8921195268631}\n",
      "Losses {'ner': 221.85793352127075}\n",
      "Losses {'ner': 227.4035460948944}\n",
      "Losses {'ner': 208.29512524604797}\n",
      "Losses {'ner': 213.41710352897644}\n",
      "Losses {'ner': 223.69528913497925}\n",
      "Losses {'ner': 213.07961589097977}\n",
      "Losses {'ner': 213.1186022758484}\n",
      "Losses {'ner': 225.31793129444122}\n",
      "Losses {'ner': 214.17570888996124}\n",
      "Losses {'ner': 213.66760659217834}\n",
      "Losses {'ner': 205.54975432157516}\n",
      "Losses {'ner': 218.3624677658081}\n",
      "Losses {'ner': 236.6382555961609}\n",
      "Losses {'ner': 228.1580959558487}\n",
      "Losses {'ner': 234.1120467185974}\n",
      "Losses {'ner': 229.37376284599304}\n",
      "Losses {'ner': 222.48266923427582}\n",
      "Losses {'ner': 225.74272775650024}\n",
      "Losses {'ner': 220.45494842529297}\n",
      "Losses {'ner': 224.06686758995056}\n",
      "Losses {'ner': 215.2231204509735}\n",
      "Losses {'ner': 206.19009387493134}\n",
      "Losses {'ner': 230.52929818630219}\n",
      "Losses {'ner': 227.54796993732452}\n",
      "Losses {'ner': 217.43470549583435}\n",
      "Losses {'ner': 224.86373269557953}\n",
      "Losses {'ner': 218.02460378408432}\n",
      "Losses {'ner': 210.34443151950836}\n",
      "Losses {'ner': 216.1765899658203}\n",
      "Losses {'ner': 213.60358715057373}\n",
      "Losses {'ner': 216.9412682056427}\n",
      "Losses {'ner': 215.95535099506378}\n",
      "Losses {'ner': 225.81547838449478}\n",
      "Losses {'ner': 220.84729087352753}\n",
      "Losses {'ner': 210.91957473754883}\n",
      "Losses {'ner': 224.57508277893066}\n",
      "Losses {'ner': 221.33657360076904}\n",
      "Losses {'ner': 228.00457155704498}\n",
      "Losses {'ner': 216.14848017692566}\n",
      "Losses {'ner': 223.7163064479828}\n",
      "Losses {'ner': 225.57256484031677}\n",
      "Losses {'ner': 227.65561592578888}\n",
      "Losses {'ner': 222.85964906215668}\n",
      "Losses {'ner': 237.9178626537323}\n",
      "Losses {'ner': 217.79623174667358}\n",
      "Losses {'ner': 216.94323205947876}\n",
      "Losses {'ner': 228.52466678619385}\n",
      "Losses {'ner': 229.88658392429352}\n",
      "Losses {'ner': 224.90158689022064}\n",
      "Losses {'ner': 220.2712286710739}\n",
      "Losses {'ner': 221.54907524585724}\n",
      "Losses {'ner': 225.23130178451538}\n",
      "Losses {'ner': 234.48440766334534}\n",
      "Losses {'ner': 229.88874101638794}\n",
      "Losses {'ner': 221.1135058403015}\n",
      "Losses {'ner': 197.71720480918884}\n",
      "Losses {'ner': 221.39548349380493}\n",
      "Losses {'ner': 218.85398149490356}\n",
      "Losses {'ner': 220.52474808692932}\n",
      "Losses {'ner': 221.0352485179901}\n",
      "Losses {'ner': 197.66532278060913}\n",
      "Losses {'ner': 224.33012235164642}\n",
      "Losses {'ner': 212.23305642604828}\n",
      "Losses {'ner': 218.9881453514099}\n",
      "Losses {'ner': 206.27466320991516}\n",
      "Losses {'ner': 230.07642608880997}\n",
      "Losses {'ner': 227.1453948020935}\n",
      "Losses {'ner': 219.67566812038422}\n",
      "Losses {'ner': 233.19076723605394}\n",
      "Losses {'ner': 216.06697535514832}\n",
      "Losses {'ner': 206.6433174610138}\n",
      "Losses {'ner': 221.70047175884247}\n",
      "Losses {'ner': 233.05878794193268}\n",
      "Losses {'ner': 224.7596926689148}\n",
      "Losses {'ner': 221.2660266160965}\n",
      "Losses {'ner': 205.08191138505936}\n",
      "Losses {'ner': 224.0840287208557}\n",
      "Losses {'ner': 215.64516305923462}\n",
      "Losses {'ner': 228.7245011329651}\n",
      "Losses {'ner': 243.1926760673523}\n",
      "Losses {'ner': 213.46277758479118}\n",
      "Losses {'ner': 219.56145691871643}\n",
      "Losses {'ner': 212.3452297449112}\n",
      "Losses {'ner': 221.39489793777466}\n",
      "Losses {'ner': 198.43953919410706}\n",
      "Losses {'ner': 216.35895954817533}\n",
      "Losses {'ner': 230.057461977005}\n",
      "Losses {'ner': 213.43238079547882}\n",
      "Losses {'ner': 199.92712903022766}\n",
      "Losses {'ner': 225.40743613243103}\n",
      "Losses {'ner': 224.2011262178421}\n",
      "Losses {'ner': 211.92556244134903}\n",
      "Losses {'ner': 221.3592073917389}\n",
      "Losses {'ner': 218.17778658866882}\n",
      "Losses {'ner': 219.57589292526245}\n",
      "Losses {'ner': 229.39809203147888}\n",
      "Losses {'ner': 220.92146515846252}\n",
      "Losses {'ner': 221.920947432518}\n",
      "Losses {'ner': 222.09745144844055}\n",
      "Losses {'ner': 214.51295518875122}\n",
      "Losses {'ner': 238.2373976111412}\n",
      "Losses {'ner': 214.4640200138092}\n",
      "Losses {'ner': 214.5680240392685}\n",
      "Losses {'ner': 215.48256635665894}\n",
      "Losses {'ner': 236.77787494659424}\n",
      "Losses {'ner': 238.88695633411407}\n",
      "Losses {'ner': 222.5149326324463}\n",
      "Losses {'ner': 222.38150215148926}\n",
      "Losses {'ner': 229.08500027656555}\n",
      "Losses {'ner': 222.95183491706848}\n",
      "Losses {'ner': 220.2005844116211}\n",
      "Losses {'ner': 228.05837321281433}\n",
      "Losses {'ner': 204.5641804933548}\n",
      "Losses {'ner': 221.72722744941711}\n",
      "Losses {'ner': 222.21970856189728}\n",
      "Losses {'ner': 218.3259139060974}\n",
      "Losses {'ner': 209.06318795681}\n",
      "Losses {'ner': 223.14697670936584}\n",
      "Losses {'ner': 227.6876545548439}\n",
      "Losses {'ner': 221.2789363861084}\n",
      "Losses {'ner': 236.7634789943695}\n",
      "Losses {'ner': 209.12371921539307}\n",
      "Losses {'ner': 224.99802708625793}\n",
      "Losses {'ner': 227.9157738685608}\n",
      "Losses {'ner': 217.9094421863556}\n",
      "Losses {'ner': 222.1872661113739}\n",
      "Losses {'ner': 210.92249941825867}\n",
      "Losses {'ner': 225.00518655776978}\n",
      "Losses {'ner': 219.8143824338913}\n",
      "Losses {'ner': 216.246191740036}\n",
      "Losses {'ner': 206.5666799545288}\n",
      "Losses {'ner': 204.53613674640656}\n",
      "Losses {'ner': 219.6089608669281}\n",
      "Losses {'ner': 210.2100579738617}\n",
      "Losses {'ner': 220.3213300704956}\n",
      "Losses {'ner': 226.07578963041306}\n",
      "Losses {'ner': 221.51738703250885}\n",
      "Losses {'ner': 210.9410154223442}\n",
      "Losses {'ner': 228.0252013206482}\n",
      "Losses {'ner': 235.70379102230072}\n",
      "Losses {'ner': 222.54188418388367}\n",
      "Losses {'ner': 220.59155201911926}\n",
      "Losses {'ner': 226.40420603752136}\n",
      "Losses {'ner': 223.43849861621857}\n",
      "Losses {'ner': 232.05407989025116}\n",
      "Losses {'ner': 220.9435703754425}\n",
      "Losses {'ner': 217.05499053001404}\n",
      "Losses {'ner': 222.3851387500763}\n",
      "Losses {'ner': 224.65684461593628}\n",
      "Losses {'ner': 225.49469530582428}\n",
      "Losses {'ner': 218.8176634311676}\n",
      "Losses {'ner': 239.8288961648941}\n",
      "Losses {'ner': 222.05629754066467}\n",
      "Losses {'ner': 213.86354780197144}\n",
      "Losses {'ner': 219.96705436706543}\n",
      "Losses {'ner': 220.97584533691406}\n",
      "Losses {'ner': 222.25982236862183}\n",
      "Losses {'ner': 226.83086156845093}\n",
      "Losses {'ner': 219.30733680725098}\n",
      "Losses {'ner': 221.96242594718933}\n",
      "Losses {'ner': 224.69104051589966}\n",
      "Losses {'ner': 215.87022137641907}\n",
      "Losses {'ner': 227.71218466758728}\n",
      "Losses {'ner': 215.95555675029755}\n",
      "Losses {'ner': 214.5823426246643}\n",
      "Losses {'ner': 217.19070386886597}\n",
      "Losses {'ner': 230.21788811683655}\n",
      "Losses {'ner': 222.91182446479797}\n",
      "Losses {'ner': 234.29712557792664}\n",
      "Losses {'ner': 212.7590161561966}\n",
      "Losses {'ner': 220.11725878715515}\n",
      "Losses {'ner': 224.91139268875122}\n",
      "Losses {'ner': 236.8663318157196}\n",
      "Losses {'ner': 222.7876100540161}\n",
      "Losses {'ner': 220.61591124534607}\n",
      "Losses {'ner': 226.72067999839783}\n",
      "Losses {'ner': 228.34184622764587}\n",
      "Losses {'ner': 227.54852187633514}\n",
      "Losses {'ner': 222.35465574264526}\n",
      "Losses {'ner': 216.9269073009491}\n",
      "Losses {'ner': 225.31839334964752}\n",
      "Losses {'ner': 216.78588008880615}\n",
      "Losses {'ner': 224.6491676568985}\n",
      "Losses {'ner': 213.90530574321747}\n",
      "Losses {'ner': 231.03015220165253}\n",
      "Losses {'ner': 232.9329903125763}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 228.0916644334793}\n",
      "Losses {'ner': 218.7370307445526}\n",
      "Losses {'ner': 221.99818420410156}\n",
      "Losses {'ner': 229.736674785614}\n",
      "Losses {'ner': 223.2610683441162}\n",
      "Losses {'ner': 225.63313841819763}\n",
      "Losses {'ner': 235.09440112113953}\n",
      "Losses {'ner': 231.8057541847229}\n",
      "Losses {'ner': 223.67653250694275}\n",
      "Losses {'ner': 234.12755131721497}\n",
      "Losses {'ner': 224.78094232082367}\n",
      "Losses {'ner': 226.91138446331024}\n",
      "Losses {'ner': 216.1558632850647}\n",
      "Losses {'ner': 219.21481049060822}\n",
      "Losses {'ner': 218.8179793357849}\n",
      "Losses {'ner': 230.92865633964539}\n",
      "Losses {'ner': 218.60689298808575}\n",
      "Losses {'ner': 235.37648820877075}\n",
      "Losses {'ner': 216.17192780971527}\n",
      "Losses {'ner': 227.10095977783203}\n",
      "Losses {'ner': 222.75739908218384}\n",
      "Losses {'ner': 225.5548918247223}\n",
      "Losses {'ner': 231.6746380329132}\n",
      "Losses {'ner': 217.83766508102417}\n",
      "Losses {'ner': 219.36157047748566}\n",
      "Losses {'ner': 226.20853734016418}\n",
      "Losses {'ner': 228.89464020729065}\n",
      "Losses {'ner': 224.2955572605133}\n",
      "Losses {'ner': 224.2631494998932}\n",
      "Losses {'ner': 232.44184279441833}\n",
      "Losses {'ner': 224.62089294195175}\n",
      "Losses {'ner': 217.8143676519394}\n",
      "Losses {'ner': 230.32470703125}\n",
      "Losses {'ner': 224.7078137397766}\n",
      "Losses {'ner': 227.26258635520935}\n",
      "Losses {'ner': 231.24278616905212}\n",
      "Losses {'ner': 220.1131843328476}\n",
      "Losses {'ner': 225.3618540763855}\n",
      "Losses {'ner': 227.7485395669937}\n",
      "Losses {'ner': 209.56454694271088}\n",
      "Losses {'ner': 222.1773431301117}\n",
      "Losses {'ner': 218.78074979782104}\n",
      "Losses {'ner': 233.05863499641418}\n",
      "Losses {'ner': 223.88096714019775}\n",
      "Losses {'ner': 218.34626162052155}\n",
      "Losses {'ner': 214.614608168602}\n",
      "Losses {'ner': 218.56812477111816}\n",
      "Losses {'ner': 208.01642107963562}\n",
      "Losses {'ner': 214.3292219042778}\n",
      "Losses {'ner': 223.47778344154358}\n",
      "Losses {'ner': 214.48286485671997}\n",
      "Losses {'ner': 221.72645592689514}\n",
      "Losses {'ner': 221.7976953983307}\n",
      "Losses {'ner': 218.70247745513916}\n",
      "Losses {'ner': 210.68250632286072}\n",
      "Losses {'ner': 226.31470239162445}\n",
      "Losses {'ner': 222.9244431257248}\n",
      "Losses {'ner': 228.38362497091293}\n",
      "Losses {'ner': 229.28288114070892}\n",
      "Losses {'ner': 221.0941594839096}\n",
      "Losses {'ner': 216.2205626964569}\n",
      "Losses {'ner': 224.55257511138916}\n",
      "Losses {'ner': 207.588568598032}\n",
      "Losses {'ner': 224.0846517086029}\n",
      "Losses {'ner': 221.336530148983}\n",
      "Losses {'ner': 218.5151093006134}\n",
      "Losses {'ner': 217.63925325870514}\n",
      "Losses {'ner': 212.5614396929741}\n",
      "Losses {'ner': 219.77176022529602}\n",
      "Losses {'ner': 223.4263617992401}\n",
      "Losses {'ner': 215.26209008693695}\n",
      "Losses {'ner': 227.42994511127472}\n",
      "Losses {'ner': 228.31408441066742}\n",
      "Losses {'ner': 227.58738708496094}\n",
      "Losses {'ner': 220.46462363004684}\n",
      "Losses {'ner': 219.54219341278076}\n",
      "Losses {'ner': 227.17531967163086}\n",
      "Losses {'ner': 218.64122092723846}\n",
      "Losses {'ner': 214.25124764442444}\n",
      "Losses {'ner': 221.96240186691284}\n",
      "Losses {'ner': 232.0302803516388}\n",
      "Losses {'ner': 224.72182512283325}\n",
      "Losses {'ner': 221.40171599388123}\n",
      "Losses {'ner': 228.22213768959045}\n",
      "Losses {'ner': 221.87113499641418}\n",
      "Losses {'ner': 230.54363524913788}\n",
      "Losses {'ner': 218.97546541690826}\n",
      "Losses {'ner': 212.38071608543396}\n",
      "Losses {'ner': 218.99720633029938}\n",
      "Losses {'ner': 214.2711787223816}\n",
      "Losses {'ner': 224.19788670539856}\n",
      "Losses {'ner': 234.6582419872284}\n",
      "Losses {'ner': 226.35400772094727}\n",
      "Losses {'ner': 215.4810849428177}\n",
      "Losses {'ner': 225.23153853416443}\n",
      "Losses {'ner': 223.85086750984192}\n",
      "Losses {'ner': 222.485009431839}\n",
      "Losses {'ner': 217.9897322654724}\n",
      "Losses {'ner': 227.92094087600708}\n",
      "Losses {'ner': 236.91390180587769}\n",
      "Losses {'ner': 218.84420764446259}\n",
      "Losses {'ner': 222.5298655629158}\n",
      "Losses {'ner': 220.6997011899948}\n",
      "Losses {'ner': 232.54385614395142}\n",
      "Losses {'ner': 233.37316304445267}\n",
      "Losses {'ner': 219.42402470111847}\n",
      "Losses {'ner': 232.94421100616455}\n",
      "Losses {'ner': 228.814284324646}\n",
      "Losses {'ner': 220.47699511051178}\n",
      "Losses {'ner': 223.40633258223534}\n",
      "Losses {'ner': 219.72970414161682}\n",
      "Losses {'ner': 226.32668960094452}\n",
      "Losses {'ner': 216.68764728307724}\n",
      "Losses {'ner': 220.37521767616272}\n",
      "Losses {'ner': 239.6520800590515}\n",
      "Losses {'ner': 210.0703706741333}\n",
      "Losses {'ner': 225.1235761642456}\n",
      "Losses {'ner': 227.92542046308517}\n",
      "Losses {'ner': 231.33774495124817}\n",
      "Losses {'ner': 220.19643712043762}\n",
      "Losses {'ner': 220.26419138908386}\n",
      "Losses {'ner': 222.6162142753601}\n",
      "Losses {'ner': 209.52000534534454}\n",
      "Losses {'ner': 215.46807026863098}\n",
      "Losses {'ner': 234.35094499588013}\n",
      "Losses {'ner': 221.66166698932648}\n",
      "Losses {'ner': 228.44658756256104}\n",
      "Losses {'ner': 223.770401597023}\n",
      "Losses {'ner': 217.21410620212555}\n",
      "Losses {'ner': 215.72367143630981}\n",
      "Losses {'ner': 230.53862261772156}\n",
      "Losses {'ner': 228.84936439990997}\n",
      "Losses {'ner': 220.94491982460022}\n",
      "Losses {'ner': 232.66419792175293}\n",
      "Losses {'ner': 212.05310654640198}\n",
      "Losses {'ner': 217.87947952747345}\n",
      "Losses {'ner': 220.13039374351501}\n",
      "Losses {'ner': 220.21690618991852}\n",
      "Losses {'ner': 229.34746837615967}\n",
      "Losses {'ner': 228.82712054252625}\n",
      "Losses {'ner': 216.31047987937927}\n",
      "Losses {'ner': 217.12336575984955}\n",
      "Losses {'ner': 226.59443426132202}\n",
      "Losses {'ner': 214.55656170845032}\n",
      "Losses {'ner': 216.11741715669632}\n",
      "Losses {'ner': 228.00312459468842}\n",
      "Losses {'ner': 223.58965945243835}\n",
      "Losses {'ner': 224.26065754890442}\n",
      "Losses {'ner': 219.77855825424194}\n",
      "Losses {'ner': 230.9723436832428}\n",
      "Losses {'ner': 232.6148281097412}\n",
      "Losses {'ner': 227.40172934532166}\n",
      "Losses {'ner': 210.37437105178833}\n",
      "Losses {'ner': 222.2222023010254}\n",
      "Losses {'ner': 218.32038187980652}\n",
      "Losses {'ner': 234.3680620789528}\n",
      "Losses {'ner': 229.74344754219055}\n",
      "Losses {'ner': 208.39783704280853}\n",
      "Losses {'ner': 213.8476390838623}\n",
      "Losses {'ner': 217.96350318193436}\n",
      "Losses {'ner': 224.75097262859344}\n",
      "Losses {'ner': 219.51918733119965}\n",
      "Losses {'ner': 217.36477613449097}\n",
      "Losses {'ner': 209.9681260585785}\n",
      "Losses {'ner': 206.50780177116394}\n",
      "Losses {'ner': 223.98164868354797}\n",
      "Losses {'ner': 237.92411541938782}\n",
      "Losses {'ner': 230.9706710577011}\n",
      "Losses {'ner': 200.44716691970825}\n",
      "Losses {'ner': 222.95904171466827}\n",
      "Losses {'ner': 218.40299677848816}\n",
      "Losses {'ner': 214.04955101013184}\n",
      "Losses {'ner': 228.7079268693924}\n",
      "Losses {'ner': 213.53973805904388}\n",
      "Losses {'ner': 226.76955032348633}\n",
      "Losses {'ner': 227.6939357817173}\n",
      "Losses {'ner': 234.88575339317322}\n",
      "Losses {'ner': 209.55052852630615}\n",
      "Losses {'ner': 227.44889426231384}\n",
      "Losses {'ner': 228.39125728607178}\n",
      "Losses {'ner': 217.72532165050507}\n",
      "Losses {'ner': 218.74503684043884}\n",
      "Losses {'ner': 223.30693078041077}\n",
      "Losses {'ner': 225.1058223247528}\n",
      "Losses {'ner': 220.42621040344238}\n",
      "Losses {'ner': 222.1690878868103}\n",
      "Losses {'ner': 233.62927520275116}\n",
      "Losses {'ner': 227.84588599205017}\n",
      "Losses {'ner': 230.38325905799866}\n",
      "Losses {'ner': 222.2766752243042}\n",
      "Losses {'ner': 219.27863228321075}\n",
      "Losses {'ner': 234.09741616249084}\n",
      "Losses {'ner': 232.72801542282104}\n",
      "Losses {'ner': 224.74043548107147}\n",
      "Losses {'ner': 231.56275308132172}\n",
      "Losses {'ner': 218.4951496720314}\n",
      "Losses {'ner': 229.97163248062134}\n",
      "Losses {'ner': 207.01603519916534}\n",
      "Losses {'ner': 226.73742604255676}\n",
      "Losses {'ner': 211.69366031885147}\n",
      "Losses {'ner': 214.10046243667603}\n",
      "Losses {'ner': 199.0610008239746}\n",
      "Losses {'ner': 220.53787052631378}\n",
      "Losses {'ner': 221.93626952171326}\n",
      "Losses {'ner': 216.24266254901886}\n",
      "Losses {'ner': 221.33310997486115}\n",
      "Losses {'ner': 228.28505563735962}\n",
      "Losses {'ner': 225.89343762397766}\n",
      "Losses {'ner': 212.98853588104248}\n",
      "Losses {'ner': 226.44780439138412}\n",
      "Losses {'ner': 219.62587797641754}\n",
      "Losses {'ner': 217.16452860832214}\n",
      "Losses {'ner': 222.72706699371338}\n",
      "Losses {'ner': 215.01749730110168}\n",
      "Losses {'ner': 214.63874685764313}\n",
      "Losses {'ner': 221.70552706718445}\n",
      "Losses {'ner': 219.26352310180664}\n",
      "Losses {'ner': 221.3712441921234}\n",
      "Losses {'ner': 219.13217343389988}\n",
      "Losses {'ner': 219.28768864274025}\n",
      "Losses {'ner': 224.27272057533264}\n",
      "Losses {'ner': 226.6197054386139}\n",
      "Losses {'ner': 231.23671901226044}\n",
      "Losses {'ner': 222.5615332722664}\n",
      "Losses {'ner': 220.43612670898438}\n",
      "Losses {'ner': 218.69462394714355}\n",
      "Losses {'ner': 218.58976936340332}\n",
      "Losses {'ner': 223.83326488733292}\n",
      "Losses {'ner': 228.9242341518402}\n",
      "Losses {'ner': 222.58632111549377}\n",
      "Losses {'ner': 230.89860546588898}\n",
      "Losses {'ner': 228.31475043296814}\n",
      "Losses {'ner': 210.22799277305603}\n",
      "Losses {'ner': 228.87689864635468}\n",
      "Losses {'ner': 195.79144859313965}\n",
      "Losses {'ner': 213.73718893527985}\n",
      "Losses {'ner': 204.16968202590942}\n",
      "Losses {'ner': 221.57589596509933}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 202.74308902025223}\n",
      "Losses {'ner': 207.17929136753082}\n",
      "Losses {'ner': 223.85659909248352}\n",
      "Losses {'ner': 233.11494851112366}\n",
      "Losses {'ner': 217.5611046552658}\n",
      "Losses {'ner': 228.65242421627045}\n",
      "Losses {'ner': 217.22935211658478}\n",
      "Losses {'ner': 221.89943730831146}\n",
      "Losses {'ner': 210.1746711730957}\n",
      "Losses {'ner': 219.42660880088806}\n",
      "Losses {'ner': 230.44040989875793}\n",
      "Losses {'ner': 219.49887120723724}\n",
      "Losses {'ner': 224.24582254886627}\n",
      "Losses {'ner': 219.84455060958862}\n",
      "Losses {'ner': 220.15121936798096}\n",
      "Losses {'ner': 222.7286732196808}\n",
      "Losses {'ner': 224.3525092601776}\n",
      "Losses {'ner': 219.4525020122528}\n",
      "Losses {'ner': 215.08594650030136}\n",
      "Losses {'ner': 229.4779823422432}\n",
      "Losses {'ner': 221.6293981075287}\n",
      "Losses {'ner': 221.35967049002647}\n",
      "Losses {'ner': 230.06424057483673}\n",
      "Losses {'ner': 228.1003077030182}\n",
      "Losses {'ner': 211.03849577903748}\n",
      "Losses {'ner': 224.04109382629395}\n",
      "Losses {'ner': 210.93128514289856}\n",
      "Losses {'ner': 208.12825083732605}\n",
      "Losses {'ner': 214.75421047210693}\n",
      "Losses {'ner': 211.6534218788147}\n",
      "Losses {'ner': 216.27977865934372}\n",
      "Losses {'ner': 225.36905282735825}\n",
      "Losses {'ner': 215.49735069274902}\n",
      "Losses {'ner': 229.7621911764145}\n",
      "Losses {'ner': 223.3234143257141}\n",
      "Losses {'ner': 225.43191719055176}\n",
      "Losses {'ner': 227.49442315101624}\n",
      "Losses {'ner': 202.50521683692932}\n",
      "Losses {'ner': 230.59089422225952}\n",
      "Losses {'ner': 224.10495138168335}\n",
      "Losses {'ner': 228.01063704490662}\n",
      "Losses {'ner': 220.05935311317444}\n",
      "Losses {'ner': 237.05541276931763}\n",
      "Losses {'ner': 215.02124899625778}\n",
      "Losses {'ner': 221.04295444488525}\n",
      "Losses {'ner': 220.2199788093567}\n",
      "Losses {'ner': 207.8745995759964}\n",
      "Losses {'ner': 221.49248641729355}\n",
      "Losses {'ner': 222.0375964641571}\n",
      "Losses {'ner': 218.70673155784607}\n",
      "Losses {'ner': 212.22405982017517}\n",
      "Losses {'ner': 229.15091371536255}\n",
      "Losses {'ner': 222.51537215709686}\n",
      "Losses {'ner': 232.28229475021362}\n",
      "Losses {'ner': 224.99004244804382}\n",
      "Losses {'ner': 204.68838024139404}\n",
      "Losses {'ner': 213.98960852622986}\n",
      "Losses {'ner': 224.8334038257599}\n",
      "Losses {'ner': 211.24835109710693}\n",
      "Losses {'ner': 227.8321876525879}\n",
      "Losses {'ner': 226.8813853263855}\n",
      "Losses {'ner': 215.53789138793945}\n",
      "Losses {'ner': 214.5016884803772}\n",
      "Losses {'ner': 230.0370535850525}\n",
      "Losses {'ner': 219.7890932559967}\n",
      "Losses {'ner': 228.0675208568573}\n",
      "Losses {'ner': 212.976487159729}\n",
      "Losses {'ner': 227.5090320110321}\n",
      "Losses {'ner': 223.53988313674927}\n",
      "Losses {'ner': 212.2564743757248}\n",
      "Losses {'ner': 214.55272299051285}\n",
      "Losses {'ner': 215.60029470920563}\n",
      "Losses {'ner': 232.95653879642487}\n",
      "Losses {'ner': 222.83012676239014}\n",
      "Losses {'ner': 220.29410415887833}\n",
      "Losses {'ner': 217.50683677196503}\n",
      "Losses {'ner': 219.11206901073456}\n",
      "Losses {'ner': 214.59222853183746}\n",
      "Losses {'ner': 239.1479948759079}\n",
      "Losses {'ner': 214.92219948768616}\n",
      "Losses {'ner': 211.5122582912445}\n",
      "Losses {'ner': 212.98623287677765}\n",
      "Losses {'ner': 222.3349860906601}\n",
      "Losses {'ner': 218.45647609233856}\n",
      "Losses {'ner': 218.47359770536423}\n",
      "Losses {'ner': 212.1229054927826}\n",
      "Losses {'ner': 215.48517495393753}\n",
      "Losses {'ner': 207.6717119216919}\n",
      "Losses {'ner': 225.81956458091736}\n",
      "Losses {'ner': 221.6361837387085}\n",
      "Losses {'ner': 224.27460408210754}\n",
      "Losses {'ner': 232.66666519641876}\n",
      "Losses {'ner': 217.54878294467926}\n",
      "Losses {'ner': 214.38257718086243}\n",
      "Losses {'ner': 220.7940616607666}\n",
      "Losses {'ner': 214.99953961372375}\n",
      "Losses {'ner': 229.92256665229797}\n",
      "Losses {'ner': 222.71748661994934}\n",
      "Losses {'ner': 224.59960174560547}\n",
      "Losses {'ner': 204.57972311973572}\n",
      "Losses {'ner': 217.2197642326355}\n",
      "Losses {'ner': 211.06839883327484}\n",
      "Losses {'ner': 213.00915694236755}\n",
      "Losses {'ner': 212.96823120117188}\n",
      "Losses {'ner': 224.67980620265007}\n",
      "Losses {'ner': 221.77884089946747}\n",
      "Losses {'ner': 235.75930523872375}\n",
      "Losses {'ner': 219.56555807590485}\n",
      "Losses {'ner': 230.0183811187744}\n",
      "Losses {'ner': 222.79643273353577}\n",
      "Losses {'ner': 212.31154453754425}\n",
      "Losses {'ner': 215.2450419664383}\n",
      "Losses {'ner': 219.78228199481964}\n",
      "Losses {'ner': 219.79741835594177}\n",
      "Losses {'ner': 216.4540577530861}\n",
      "Losses {'ner': 215.9858993291855}\n",
      "Losses {'ner': 205.41582131385803}\n",
      "Losses {'ner': 209.12891745567322}\n",
      "Losses {'ner': 222.16582417488098}\n",
      "Losses {'ner': 223.54556465148926}\n",
      "Losses {'ner': 213.20746290683746}\n",
      "Losses {'ner': 227.23203694820404}\n",
      "Losses {'ner': 227.60206866264343}\n",
      "Losses {'ner': 222.06478410959244}\n",
      "Losses {'ner': 225.9256820678711}\n",
      "Losses {'ner': 216.08323633670807}\n",
      "Losses {'ner': 214.04001355171204}\n",
      "Losses {'ner': 218.92768293619156}\n",
      "Losses {'ner': 221.86887514591217}\n",
      "Losses {'ner': 219.28583604097366}\n",
      "Losses {'ner': 219.29226779937744}\n",
      "Losses {'ner': 216.98381316661835}\n",
      "Losses {'ner': 229.11033058166504}\n",
      "Losses {'ner': 212.659609913826}\n",
      "Losses {'ner': 211.62409281730652}\n",
      "Losses {'ner': 215.9827275276184}\n",
      "Losses {'ner': 234.44778728485107}\n",
      "Losses {'ner': 203.16578710079193}\n",
      "Losses {'ner': 219.90814048051834}\n",
      "Losses {'ner': 213.45554876327515}\n",
      "Losses {'ner': 214.82180845737457}\n",
      "Losses {'ner': 225.3243052959442}\n",
      "Losses {'ner': 214.14209127426147}\n",
      "Losses {'ner': 216.4678297638893}\n",
      "Losses {'ner': 220.60069227218628}\n",
      "Losses {'ner': 229.71594405174255}\n",
      "Losses {'ner': 235.5115041732788}\n",
      "Losses {'ner': 219.81455028057098}\n",
      "Losses {'ner': 221.67477297782898}\n",
      "Losses {'ner': 224.5547013282776}\n",
      "Losses {'ner': 221.03582382202148}\n",
      "Losses {'ner': 218.1403752565384}\n",
      "Losses {'ner': 227.81350445747375}\n",
      "Losses {'ner': 216.80220210552216}\n",
      "Losses {'ner': 228.67513871192932}\n",
      "Losses {'ner': 204.32353949546814}\n",
      "Losses {'ner': 217.86380749940872}\n",
      "Losses {'ner': 220.5417994260788}\n",
      "Losses {'ner': 211.6714165210724}\n",
      "Losses {'ner': 217.59943318367004}\n",
      "Losses {'ner': 230.76241612434387}\n",
      "Losses {'ner': 214.57881200313568}\n",
      "Losses {'ner': 214.7235586643219}\n",
      "Losses {'ner': 225.73872065544128}\n",
      "Losses {'ner': 220.51132774353027}\n",
      "Losses {'ner': 208.66426873207092}\n",
      "Losses {'ner': 228.6111330986023}\n",
      "Losses {'ner': 221.98503828048706}\n",
      "Losses {'ner': 226.36092233657837}\n",
      "Losses {'ner': 229.70332026481628}\n",
      "Losses {'ner': 222.49665546417236}\n",
      "Losses {'ner': 219.58477091789246}\n",
      "Losses {'ner': 222.50444996356964}\n",
      "Losses {'ner': 224.84635186195374}\n",
      "Losses {'ner': 214.3937168121338}\n",
      "Losses {'ner': 236.05386424064636}\n",
      "Losses {'ner': 227.83141946792603}\n",
      "Losses {'ner': 222.8507221341133}\n",
      "Losses {'ner': 221.69978606700897}\n",
      "Losses {'ner': 223.0403949022293}\n",
      "Losses {'ner': 218.34376859664917}\n",
      "Losses {'ner': 238.8109586238861}\n",
      "Losses {'ner': 220.41629254817963}\n",
      "Losses {'ner': 220.1870813369751}\n",
      "Losses {'ner': 222.9478816986084}\n",
      "Losses {'ner': 214.19141054153442}\n",
      "Losses {'ner': 213.64442437887192}\n",
      "Losses {'ner': 228.63740849494934}\n",
      "Losses {'ner': 227.07866072654724}\n",
      "Losses {'ner': 225.34619867801666}\n",
      "Losses {'ner': 218.990982234478}\n",
      "Losses {'ner': 231.99074292182922}\n",
      "Losses {'ner': 221.67785930633545}\n",
      "Losses {'ner': 227.5630885362625}\n",
      "Losses {'ner': 216.33776688575745}\n",
      "Losses {'ner': 230.31465542316437}\n",
      "Losses {'ner': 224.68903231620789}\n",
      "Losses {'ner': 210.1129047870636}\n",
      "Losses {'ner': 222.62754464149475}\n",
      "Losses {'ner': 212.49721109867096}\n",
      "Losses {'ner': 217.0886709690094}\n",
      "Losses {'ner': 217.62048137187958}\n",
      "Losses {'ner': 230.0512821674347}\n",
      "Losses {'ner': 225.22736823558807}\n",
      "Losses {'ner': 229.14772045612335}\n",
      "Losses {'ner': 217.73153948783875}\n",
      "Losses {'ner': 225.50115060806274}\n",
      "Losses {'ner': 224.55322134494781}\n",
      "Losses {'ner': 233.31434309482574}\n",
      "Losses {'ner': 223.1913743019104}\n",
      "Losses {'ner': 213.1857522726059}\n",
      "Losses {'ner': 215.70033431053162}\n",
      "Losses {'ner': 211.9802668094635}\n",
      "Losses {'ner': 222.40752041339874}\n",
      "Losses {'ner': 212.31998777389526}\n",
      "Losses {'ner': 217.29956376552582}\n",
      "Losses {'ner': 215.86924040317535}\n",
      "Losses {'ner': 193.1285127401352}\n",
      "Losses {'ner': 210.1209135055542}\n",
      "Losses {'ner': 218.97238993644714}\n",
      "Losses {'ner': 230.19946312904358}\n",
      "Losses {'ner': 211.94232058525085}\n",
      "Losses {'ner': 224.51485800743103}\n",
      "Losses {'ner': 231.79176664352417}\n",
      "Losses {'ner': 210.10702562332153}\n",
      "Losses {'ner': 214.12569999694824}\n",
      "Losses {'ner': 226.8717629313469}\n",
      "Losses {'ner': 229.64397013187408}\n",
      "Losses {'ner': 220.13894999027252}\n",
      "Losses {'ner': 215.10231709480286}\n",
      "Losses {'ner': 203.77503895759583}\n",
      "Losses {'ner': 222.01410871744156}\n",
      "Losses {'ner': 217.23481357097626}\n",
      "Losses {'ner': 234.4801283478737}\n",
      "Losses {'ner': 216.84534215927124}\n",
      "Losses {'ner': 217.69948422908783}\n",
      "Losses {'ner': 235.90934538841248}\n",
      "Losses {'ner': 231.29332613945007}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 232.55719232559204}\n",
      "Losses {'ner': 223.69595175981522}\n",
      "Losses {'ner': 211.71673786640167}\n",
      "Losses {'ner': 228.92685461044312}\n",
      "Losses {'ner': 212.2841339185834}\n",
      "Losses {'ner': 221.08789563179016}\n",
      "Losses {'ner': 231.52896070480347}\n",
      "Losses {'ner': 216.88811492919922}\n",
      "Losses {'ner': 232.5458379983902}\n",
      "Losses {'ner': 213.67218166589737}\n",
      "Losses {'ner': 220.74275487661362}\n",
      "Losses {'ner': 219.66307818889618}\n",
      "Losses {'ner': 215.28770244121552}\n",
      "Losses {'ner': 223.92075729370117}\n",
      "Losses {'ner': 235.30355191230774}\n",
      "Losses {'ner': 206.49630063772202}\n",
      "Losses {'ner': 206.64162576198578}\n",
      "Losses {'ner': 209.32990837097168}\n",
      "Losses {'ner': 238.43283009529114}\n",
      "Losses {'ner': 215.8198745250702}\n",
      "Losses {'ner': 197.22801792621613}\n",
      "Losses {'ner': 208.4958963394165}\n",
      "Losses {'ner': 211.2001303434372}\n",
      "Losses {'ner': 223.98632061481476}\n",
      "Losses {'ner': 230.09865701198578}\n",
      "Losses {'ner': 220.85997188091278}\n",
      "Losses {'ner': 236.75334203243256}\n",
      "Losses {'ner': 227.0710619688034}\n",
      "Losses {'ner': 225.19994950294495}\n",
      "Losses {'ner': 239.52515745162964}\n",
      "Losses {'ner': 226.24619352817535}\n",
      "Losses {'ner': 213.5822970867157}\n",
      "Losses {'ner': 216.24108010530472}\n",
      "Losses {'ner': 225.76720356941223}\n",
      "Losses {'ner': 215.69559025764465}\n",
      "Losses {'ner': 229.73178267478943}\n",
      "Losses {'ner': 225.34577417373657}\n",
      "Losses {'ner': 224.78852367401123}\n",
      "Losses {'ner': 221.75504851341248}\n",
      "Losses {'ner': 227.39060044288635}\n",
      "Losses {'ner': 222.56329464912415}\n",
      "Losses {'ner': 220.43214058876038}\n",
      "Losses {'ner': 223.22079813480377}\n",
      "Losses {'ner': 224.43209648132324}\n",
      "Losses {'ner': 226.12872613966465}\n",
      "Losses {'ner': 234.6735224723816}\n",
      "Losses {'ner': 233.77899956703186}\n",
      "Losses {'ner': 230.54782247543335}\n",
      "Losses {'ner': 230.86066699028015}\n",
      "Losses {'ner': 219.91844248771667}\n",
      "Losses {'ner': 219.58564978837967}\n",
      "Losses {'ner': 233.9290453195572}\n",
      "Losses {'ner': 222.4102109670639}\n",
      "Losses {'ner': 223.67410051822662}\n",
      "Losses {'ner': 225.21000409126282}\n",
      "Losses {'ner': 215.0456645488739}\n",
      "Losses {'ner': 237.3744466304779}\n",
      "Losses {'ner': 213.0652940273285}\n",
      "Losses {'ner': 228.24943947792053}\n",
      "Losses {'ner': 226.28991854190826}\n",
      "Losses {'ner': 230.6974776983261}\n",
      "Losses {'ner': 227.36897361278534}\n",
      "Losses {'ner': 231.3880339860916}\n",
      "Losses {'ner': 208.70315927267075}\n",
      "Losses {'ner': 211.2566378712654}\n",
      "Losses {'ner': 221.21196937561035}\n",
      "Losses {'ner': 230.30781829357147}\n",
      "Losses {'ner': 202.93520140647888}\n",
      "Losses {'ner': 242.63317918777466}\n",
      "Losses {'ner': 194.7519805431366}\n",
      "Losses {'ner': 217.92662167549133}\n",
      "Losses {'ner': 211.20785236358643}\n",
      "Losses {'ner': 221.08400332927704}\n",
      "Losses {'ner': 230.67153215408325}\n",
      "Losses {'ner': 217.12762415409088}\n",
      "Losses {'ner': 217.6547031402588}\n",
      "Losses {'ner': 220.23488199710846}\n",
      "Losses {'ner': 227.48661744594574}\n",
      "Losses {'ner': 227.7535182237625}\n",
      "Losses {'ner': 215.366069316864}\n",
      "Losses {'ner': 221.62626695632935}\n",
      "Losses {'ner': 231.13671457767487}\n",
      "Losses {'ner': 225.64871883392334}\n",
      "Losses {'ner': 223.4504297375679}\n",
      "Losses {'ner': 205.29768896102905}\n",
      "Losses {'ner': 214.30444359779358}\n",
      "Losses {'ner': 215.6115837097168}\n",
      "Losses {'ner': 223.97092628479004}\n",
      "Losses {'ner': 230.74865579605103}\n",
      "Losses {'ner': 216.37204694747925}\n",
      "Losses {'ner': 230.13394963741302}\n",
      "Losses {'ner': 216.45180141925812}\n",
      "Losses {'ner': 216.5552785396576}\n",
      "Losses {'ner': 215.78595089912415}\n",
      "Losses {'ner': 223.61514163017273}\n",
      "Losses {'ner': 222.1273248642683}\n",
      "Losses {'ner': 216.312739610672}\n",
      "Losses {'ner': 219.67945218086243}\n",
      "Losses {'ner': 214.84517216682434}\n",
      "Losses {'ner': 212.5566832125187}\n",
      "Losses {'ner': 222.41238355636597}\n",
      "Losses {'ner': 210.48104846477509}\n",
      "Losses {'ner': 221.68639206886292}\n",
      "Losses {'ner': 215.40733408927917}\n",
      "Losses {'ner': 229.64817237854004}\n",
      "Losses {'ner': 222.88312196731567}\n",
      "Losses {'ner': 223.52682876586914}\n",
      "Losses {'ner': 214.9557939171791}\n",
      "Losses {'ner': 215.72691941261292}\n",
      "Losses {'ner': 219.0752899646759}\n",
      "Losses {'ner': 223.72194528579712}\n",
      "Losses {'ner': 220.21378993988037}\n",
      "Losses {'ner': 203.2387752532959}\n",
      "Losses {'ner': 222.77686393260956}\n",
      "Losses {'ner': 213.86206674575806}\n",
      "Losses {'ner': 217.56672275066376}\n",
      "Losses {'ner': 223.6646906733513}\n",
      "Losses {'ner': 217.97922801971436}\n",
      "Losses {'ner': 216.0413790345192}\n",
      "Losses {'ner': 218.29716110229492}\n",
      "Losses {'ner': 231.36079323291779}\n",
      "Losses {'ner': 224.98300278186798}\n",
      "Losses {'ner': 212.01863741874695}\n",
      "Losses {'ner': 224.1680727005005}\n",
      "Losses {'ner': 202.809028506279}\n",
      "Losses {'ner': 227.68429827690125}\n",
      "Losses {'ner': 234.73205065727234}\n",
      "Losses {'ner': 207.7485467195511}\n",
      "Losses {'ner': 214.94489914178848}\n",
      "Losses {'ner': 223.01761412620544}\n",
      "Losses {'ner': 214.56026887893677}\n",
      "Losses {'ner': 214.79231333732605}\n",
      "Losses {'ner': 211.296040058136}\n",
      "Losses {'ner': 228.62000823020935}\n",
      "Losses {'ner': 203.91925567388535}\n",
      "Losses {'ner': 198.37490916252136}\n",
      "Losses {'ner': 212.06986355781555}\n",
      "Losses {'ner': 206.2657880783081}\n",
      "Losses {'ner': 217.49001455307007}\n",
      "Losses {'ner': 225.6667142510414}\n",
      "Losses {'ner': 215.8507729768753}\n",
      "Losses {'ner': 220.1486769914627}\n",
      "Losses {'ner': 211.96110355854034}\n",
      "Losses {'ner': 219.10643565654755}\n",
      "Losses {'ner': 229.43268930912018}\n",
      "Losses {'ner': 236.57117533683777}\n",
      "Losses {'ner': 213.35976898670197}\n",
      "Losses {'ner': 214.15293037891388}\n",
      "Losses {'ner': 223.8808138370514}\n",
      "Losses {'ner': 225.61459970474243}\n",
      "Losses {'ner': 225.3426035642624}\n",
      "Losses {'ner': 221.39026403427124}\n",
      "Losses {'ner': 236.75129747390747}\n",
      "Losses {'ner': 234.86725091934204}\n",
      "Losses {'ner': 212.3623330593109}\n",
      "Losses {'ner': 220.50846695899963}\n",
      "Losses {'ner': 220.83056151866913}\n",
      "Losses {'ner': 213.81851994991302}\n",
      "Losses {'ner': 229.8786895275116}\n",
      "Losses {'ner': 235.7885284423828}\n",
      "Losses {'ner': 219.7980604171753}\n",
      "Losses {'ner': 233.280339717865}\n",
      "Losses {'ner': 230.8228371143341}\n",
      "Losses {'ner': 222.06389260292053}\n",
      "Losses {'ner': 220.64014339447021}\n",
      "Losses {'ner': 219.366774559021}\n",
      "Losses {'ner': 210.779031842947}\n",
      "Losses {'ner': 233.18493962287903}\n",
      "Losses {'ner': 210.67013335227966}\n",
      "Losses {'ner': 222.52351999282837}\n",
      "Losses {'ner': 224.54415822029114}\n",
      "Losses {'ner': 217.7337189912796}\n",
      "Losses {'ner': 230.34786593914032}\n",
      "Losses {'ner': 213.89689886569977}\n",
      "Losses {'ner': 236.38102650642395}\n",
      "Losses {'ner': 237.09311866760254}\n",
      "Losses {'ner': 217.51275992393494}\n",
      "Losses {'ner': 215.53433179855347}\n",
      "Losses {'ner': 220.33363366127014}\n",
      "Losses {'ner': 215.81985759735107}\n",
      "Losses {'ner': 215.4000208377838}\n",
      "Losses {'ner': 220.7392628788948}\n",
      "Losses {'ner': 229.7575044631958}\n",
      "Losses {'ner': 212.4666509628296}\n",
      "Losses {'ner': 216.96862190961838}\n",
      "Losses {'ner': 212.85401833057404}\n",
      "Losses {'ner': 217.9516351222992}\n",
      "Losses {'ner': 223.57785725593567}\n",
      "Losses {'ner': 240.4370766878128}\n",
      "Losses {'ner': 231.94002556800842}\n",
      "Losses {'ner': 222.8020967245102}\n",
      "Losses {'ner': 219.77015417814255}\n",
      "Losses {'ner': 230.54586935043335}\n",
      "Losses {'ner': 214.70398545265198}\n",
      "Losses {'ner': 211.67009258270264}\n",
      "Losses {'ner': 231.64816975593567}\n",
      "Losses {'ner': 209.40893483161926}\n",
      "Losses {'ner': 214.97263598442078}\n",
      "Losses {'ner': 209.61688804626465}\n",
      "Losses {'ner': 222.35996413230896}\n",
      "Losses {'ner': 215.76343441009521}\n",
      "Losses {'ner': 218.43647158145905}\n",
      "Losses {'ner': 235.91616702079773}\n",
      "Losses {'ner': 225.63288617134094}\n",
      "Losses {'ner': 222.86732363700867}\n",
      "Losses {'ner': 215.30476784706116}\n",
      "Losses {'ner': 239.45907986164093}\n",
      "Losses {'ner': 226.58096951246262}\n",
      "Losses {'ner': 219.12547743320465}\n",
      "Losses {'ner': 227.76386046409607}\n",
      "Losses {'ner': 239.09626698493958}\n",
      "Losses {'ner': 216.52805364131927}\n",
      "Losses {'ner': 231.02061557769775}\n",
      "Losses {'ner': 217.64993107318878}\n",
      "Losses {'ner': 217.51922082901}\n",
      "Losses {'ner': 234.06291794776917}\n",
      "Losses {'ner': 224.4967532157898}\n",
      "Losses {'ner': 219.2876819372177}\n",
      "Losses {'ner': 225.2765793800354}\n",
      "Losses {'ner': 223.200523853302}\n",
      "Losses {'ner': 222.18801176548004}\n",
      "Losses {'ner': 226.84992241859436}\n",
      "Losses {'ner': 233.9606750011444}\n",
      "Losses {'ner': 213.07317328453064}\n",
      "Losses {'ner': 222.4015760421753}\n",
      "Losses {'ner': 223.10692501068115}\n",
      "Losses {'ner': 240.70685935020447}\n",
      "Losses {'ner': 220.24649214744568}\n",
      "Losses {'ner': 222.08707547187805}\n",
      "Losses {'ner': 228.92192888259888}\n",
      "Losses {'ner': 220.67835316061974}\n",
      "Losses {'ner': 218.31780874729156}\n",
      "Losses {'ner': 225.0922976732254}\n",
      "Losses {'ner': 217.76691246032715}\n",
      "Losses {'ner': 219.3299765586853}\n",
      "Losses {'ner': 215.24835681915283}\n",
      "Losses {'ner': 205.8067444562912}\n",
      "Losses {'ner': 213.90459382534027}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 230.82037770748138}\n",
      "Losses {'ner': 229.02386009693146}\n",
      "Losses {'ner': 210.0051827430725}\n",
      "Losses {'ner': 212.59625279903412}\n",
      "Losses {'ner': 223.64050245285034}\n",
      "Losses {'ner': 214.4969880580902}\n",
      "Losses {'ner': 210.27639782428741}\n",
      "Losses {'ner': 219.43906670808792}\n",
      "Losses {'ner': 222.63582491874695}\n",
      "Losses {'ner': 220.96052598953247}\n",
      "Losses {'ner': 203.61406445503235}\n",
      "Losses {'ner': 225.79093027114868}\n",
      "Losses {'ner': 218.46540927886963}\n",
      "Losses {'ner': 209.45237684249878}\n",
      "Losses {'ner': 225.23375010490417}\n",
      "Losses {'ner': 207.8753399848938}\n",
      "Losses {'ner': 218.38075280189514}\n",
      "Losses {'ner': 229.20478916168213}\n",
      "Losses {'ner': 204.6791934967041}\n",
      "Losses {'ner': 207.86816322803497}\n",
      "Losses {'ner': 222.97565817832947}\n",
      "Losses {'ner': 229.90834391117096}\n",
      "Losses {'ner': 224.37989258766174}\n",
      "Losses {'ner': 207.34675312042236}\n",
      "Losses {'ner': 228.71229934692383}\n",
      "Losses {'ner': 225.45970010757446}\n",
      "Losses {'ner': 217.59925645589828}\n",
      "Losses {'ner': 227.88512301445007}\n",
      "Losses {'ner': 235.99304461479187}\n",
      "Losses {'ner': 223.38896697759628}\n",
      "Losses {'ner': 224.90442991256714}\n",
      "Losses {'ner': 227.07743573188782}\n",
      "Losses {'ner': 222.1495317220688}\n",
      "Losses {'ner': 237.73499524593353}\n",
      "Losses {'ner': 221.34713506698608}\n",
      "Losses {'ner': 207.42288756370544}\n",
      "Losses {'ner': 210.20557421445847}\n",
      "Losses {'ner': 211.69774693250656}\n",
      "Losses {'ner': 225.34634232521057}\n",
      "Losses {'ner': 213.12193369865417}\n",
      "Losses {'ner': 216.7945412993431}\n",
      "Losses {'ner': 227.62883591651917}\n",
      "Losses {'ner': 220.16349720954895}\n",
      "Losses {'ner': 221.1118004322052}\n",
      "Losses {'ner': 225.60029101371765}\n",
      "Losses {'ner': 221.10752058029175}\n",
      "Losses {'ner': 218.75695741176605}\n",
      "Losses {'ner': 206.5256118774414}\n",
      "Losses {'ner': 202.59970831871033}\n",
      "Losses {'ner': 221.52480959892273}\n",
      "Losses {'ner': 230.27862763404846}\n",
      "Losses {'ner': 231.52328383922577}\n",
      "Losses {'ner': 206.3542103767395}\n",
      "Losses {'ner': 222.50501704216003}\n",
      "Losses {'ner': 227.40552258491516}\n",
      "Losses {'ner': 218.8395073413849}\n",
      "Losses {'ner': 209.1282341480255}\n",
      "Losses {'ner': 218.95567667484283}\n",
      "Losses {'ner': 226.43735313415527}\n",
      "Losses {'ner': 212.25164449214935}\n",
      "Losses {'ner': 209.61993741989136}\n",
      "Losses {'ner': 216.9196858406067}\n",
      "Losses {'ner': 224.59170842170715}\n",
      "Losses {'ner': 223.10739731788635}\n",
      "Losses {'ner': 213.04792141914368}\n",
      "Losses {'ner': 219.17459917068481}\n",
      "Losses {'ner': 224.36358898878098}\n",
      "Losses {'ner': 230.5287685394287}\n",
      "Losses {'ner': 218.3927993774414}\n",
      "Losses {'ner': 223.9999394416809}\n",
      "Losses {'ner': 214.781996011734}\n",
      "Losses {'ner': 224.08178567886353}\n",
      "Losses {'ner': 238.85866832733154}\n",
      "Losses {'ner': 211.91744709014893}\n",
      "Losses {'ner': 223.09823441505432}\n",
      "Losses {'ner': 223.7881714105606}\n",
      "Losses {'ner': 226.94711768627167}\n",
      "Losses {'ner': 223.1157032251358}\n",
      "Losses {'ner': 218.90203976631165}\n",
      "Losses {'ner': 220.5615587234497}\n",
      "Losses {'ner': 223.24763369560242}\n",
      "Losses {'ner': 221.32120156288147}\n",
      "Losses {'ner': 217.55895161628723}\n",
      "Losses {'ner': 219.10977160930634}\n",
      "Losses {'ner': 211.0625}\n",
      "Losses {'ner': 219.03067734837532}\n",
      "Losses {'ner': 231.2326375246048}\n",
      "Losses {'ner': 213.24541234970093}\n",
      "Losses {'ner': 215.39348483085632}\n",
      "Losses {'ner': 213.8362616300583}\n",
      "Losses {'ner': 223.21428418159485}\n",
      "Losses {'ner': 228.80821752548218}\n",
      "Losses {'ner': 228.5646779537201}\n",
      "Losses {'ner': 220.0853636264801}\n",
      "Losses {'ner': 224.2062108516693}\n",
      "Losses {'ner': 204.53954350948334}\n",
      "Losses {'ner': 217.0813463330269}\n",
      "Losses {'ner': 211.0451073050499}\n",
      "Losses {'ner': 231.77265870571136}\n",
      "Losses {'ner': 218.75450205802917}\n",
      "Losses {'ner': 216.7324184179306}\n",
      "Losses {'ner': 211.28433895111084}\n",
      "Losses {'ner': 214.79527878761292}\n",
      "Losses {'ner': 207.53201109170914}\n",
      "Losses {'ner': 209.0452384352684}\n",
      "Losses {'ner': 220.00385761260986}\n",
      "Losses {'ner': 218.78605662286282}\n",
      "Losses {'ner': 225.21264219284058}\n",
      "Losses {'ner': 239.16286540031433}\n",
      "Losses {'ner': 220.8873269557953}\n",
      "Losses {'ner': 213.5865330696106}\n",
      "Losses {'ner': 220.49380326271057}\n",
      "Losses {'ner': 216.82662630081177}\n",
      "Losses {'ner': 216.7480833530426}\n",
      "Losses {'ner': 221.06628453731537}\n",
      "Losses {'ner': 237.62207663059235}\n",
      "Losses {'ner': 235.87354052066803}\n",
      "Losses {'ner': 220.78752446174622}\n",
      "Losses {'ner': 224.88121795654297}\n",
      "Losses {'ner': 216.92852199077606}\n",
      "Losses {'ner': 207.7153196334839}\n",
      "Losses {'ner': 232.3446180820465}\n",
      "Losses {'ner': 214.44320225715637}\n",
      "Losses {'ner': 218.70896935462952}\n",
      "Losses {'ner': 228.82787013053894}\n",
      "Losses {'ner': 225.07773756980896}\n",
      "Losses {'ner': 220.51537156105042}\n",
      "Losses {'ner': 226.71943855285645}\n",
      "Losses {'ner': 218.13681769371033}\n",
      "Losses {'ner': 219.1127837896347}\n",
      "Losses {'ner': 220.31872761249542}\n",
      "Losses {'ner': 225.7809317111969}\n",
      "Losses {'ner': 218.10831379890442}\n",
      "Losses {'ner': 214.49793767929077}\n",
      "Losses {'ner': 224.81576669216156}\n",
      "Losses {'ner': 213.01876592636108}\n",
      "Losses {'ner': 226.7897710800171}\n",
      "Losses {'ner': 219.6173061132431}\n",
      "Losses {'ner': 241.8132827281952}\n",
      "Losses {'ner': 210.45392274856567}\n",
      "Losses {'ner': 213.21492505073547}\n",
      "Losses {'ner': 220.26944065093994}\n",
      "Losses {'ner': 213.00705218315125}\n",
      "Losses {'ner': 217.21542859077454}\n",
      "Losses {'ner': 222.00018668174744}\n",
      "Losses {'ner': 209.38236570358276}\n",
      "Losses {'ner': 221.94966703653336}\n",
      "Losses {'ner': 222.53937911987305}\n",
      "Losses {'ner': 224.45487487316132}\n",
      "Losses {'ner': 219.7598729133606}\n",
      "Losses {'ner': 213.8978110551834}\n",
      "Losses {'ner': 229.4301631450653}\n",
      "Losses {'ner': 209.90559768676758}\n",
      "Losses {'ner': 223.71451425552368}\n",
      "Losses {'ner': 215.34149825572968}\n",
      "Losses {'ner': 230.38843047618866}\n",
      "Losses {'ner': 219.47591519355774}\n",
      "Losses {'ner': 216.81341779232025}\n",
      "Losses {'ner': 222.77627968788147}\n",
      "Losses {'ner': 231.0821475982666}\n",
      "Losses {'ner': 215.15657901763916}\n",
      "Losses {'ner': 227.95975196361542}\n",
      "Losses {'ner': 209.11093950271606}\n",
      "Losses {'ner': 217.6671780347824}\n",
      "Losses {'ner': 209.5043909549713}\n",
      "Losses {'ner': 214.69769704341888}\n",
      "Losses {'ner': 220.99649024009705}\n",
      "Losses {'ner': 218.2651481628418}\n",
      "Losses {'ner': 213.32686507701874}\n",
      "Losses {'ner': 230.87794017791748}\n",
      "Losses {'ner': 209.4333873987198}\n",
      "Losses {'ner': 224.14224112033844}\n",
      "Losses {'ner': 225.97666692733765}\n",
      "Losses {'ner': 213.76630449295044}\n",
      "Losses {'ner': 220.1186192035675}\n",
      "Losses {'ner': 219.01840257644653}\n",
      "Losses {'ner': 217.84168195724487}\n",
      "Losses {'ner': 217.08788180351257}\n",
      "Losses {'ner': 227.30479884147644}\n",
      "Losses {'ner': 211.9697840809822}\n",
      "Losses {'ner': 216.69265341758728}\n",
      "Losses {'ner': 222.76464819908142}\n",
      "Losses {'ner': 220.51620864868164}\n",
      "Losses {'ner': 221.6804722547531}\n",
      "Losses {'ner': 224.62063837051392}\n",
      "Losses {'ner': 227.79492735862732}\n",
      "Losses {'ner': 215.32699233293533}\n",
      "Losses {'ner': 222.00864827632904}\n",
      "Losses {'ner': 223.90103387832642}\n",
      "Losses {'ner': 222.28273344039917}\n",
      "Losses {'ner': 214.62318363785744}\n",
      "Losses {'ner': 215.66479873657227}\n",
      "Losses {'ner': 228.17067563533783}\n",
      "Losses {'ner': 231.75852727890015}\n",
      "Losses {'ner': 215.49170541763306}\n",
      "Losses {'ner': 219.654347717762}\n",
      "Losses {'ner': 215.27153635025024}\n",
      "Losses {'ner': 206.26534700393677}\n",
      "Losses {'ner': 209.1142818927765}\n",
      "Losses {'ner': 217.53060227632523}\n",
      "Losses {'ner': 221.55406999588013}\n",
      "Losses {'ner': 219.56031906604767}\n",
      "Losses {'ner': 216.99105429649353}\n",
      "Losses {'ner': 222.10204356908798}\n",
      "Losses {'ner': 214.07994377613068}\n",
      "Losses {'ner': 221.02768683433533}\n",
      "Losses {'ner': 221.33764576911926}\n",
      "Losses {'ner': 216.19297397136688}\n",
      "Losses {'ner': 216.88094985485077}\n",
      "Losses {'ner': 229.7593126296997}\n",
      "Losses {'ner': 220.34217834472656}\n",
      "Losses {'ner': 218.24372220039368}\n",
      "Losses {'ner': 212.27405095100403}\n",
      "Losses {'ner': 226.5426001548767}\n",
      "Losses {'ner': 216.89587497711182}\n",
      "Losses {'ner': 208.63160932064056}\n",
      "Losses {'ner': 219.10487043857574}\n",
      "Losses {'ner': 231.07640051841736}\n",
      "Losses {'ner': 227.35986006259918}\n",
      "Losses {'ner': 220.65874361991882}\n",
      "Losses {'ner': 205.0768575668335}\n",
      "Losses {'ner': 228.82526206970215}\n",
      "Losses {'ner': 222.61016368865967}\n",
      "Losses {'ner': 217.48211550712585}\n",
      "Losses {'ner': 219.70412158966064}\n",
      "Losses {'ner': 229.51078510284424}\n",
      "Losses {'ner': 219.9828646183014}\n",
      "Losses {'ner': 226.19709134101868}\n",
      "Losses {'ner': 220.73625016212463}\n",
      "Losses {'ner': 234.98357248306274}\n",
      "Losses {'ner': 215.92403066158295}\n",
      "Losses {'ner': 215.4124948978424}\n",
      "Losses {'ner': 215.73259615898132}\n",
      "Losses {'ner': 234.05881786346436}\n",
      "Losses {'ner': 219.71326673030853}\n",
      "Losses {'ner': 223.53130269050598}\n",
      "Losses {'ner': 209.98669201135635}\n",
      "Losses {'ner': 218.209962785244}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 204.92081928253174}\n",
      "Losses {'ner': 216.33598750829697}\n",
      "Losses {'ner': 219.94475710391998}\n",
      "Losses {'ner': 225.99861252307892}\n",
      "Losses {'ner': 218.17990672588348}\n",
      "Losses {'ner': 214.17900490760803}\n",
      "Losses {'ner': 224.25011438131332}\n",
      "Losses {'ner': 215.57956099510193}\n",
      "Losses {'ner': 209.0237826704979}\n",
      "Losses {'ner': 210.4627387523651}\n",
      "Losses {'ner': 223.45434713363647}\n",
      "Losses {'ner': 211.38810467720032}\n",
      "Losses {'ner': 225.01898193359375}\n",
      "Losses {'ner': 226.11206150054932}\n",
      "Losses {'ner': 226.60079193115234}\n",
      "Losses {'ner': 218.03514552116394}\n",
      "Losses {'ner': 228.1552299261093}\n",
      "Losses {'ner': 217.52661406993866}\n",
      "Losses {'ner': 226.04176712036133}\n",
      "Losses {'ner': 217.42726814746857}\n",
      "Losses {'ner': 230.18410623073578}\n",
      "Losses {'ner': 207.65081596374512}\n",
      "Losses {'ner': 218.68481421470642}\n",
      "Losses {'ner': 213.41702044010162}\n",
      "Losses {'ner': 224.07016491889954}\n",
      "Losses {'ner': 216.77782142162323}\n",
      "Losses {'ner': 226.16156136989594}\n",
      "Losses {'ner': 236.00154292583466}\n",
      "Losses {'ner': 232.74839520454407}\n",
      "Losses {'ner': 216.00263118743896}\n",
      "Losses {'ner': 226.49039351940155}\n",
      "Losses {'ner': 232.16669943928719}\n",
      "Losses {'ner': 232.63310831785202}\n",
      "Losses {'ner': 216.3006683588028}\n",
      "Losses {'ner': 222.410702586174}\n",
      "Losses {'ner': 215.99334335327148}\n",
      "Losses {'ner': 223.50355863571167}\n",
      "Losses {'ner': 232.46645319461823}\n",
      "Losses {'ner': 214.42104411125183}\n",
      "Losses {'ner': 222.8191978931427}\n",
      "Losses {'ner': 233.9705879688263}\n",
      "Losses {'ner': 225.31322211027145}\n",
      "Losses {'ner': 219.9822268486023}\n",
      "Losses {'ner': 228.62006145715714}\n",
      "Losses {'ner': 215.74313282966614}\n",
      "Losses {'ner': 214.8143104314804}\n",
      "Losses {'ner': 232.23547339439392}\n",
      "Losses {'ner': 207.83077692985535}\n",
      "Losses {'ner': 227.58522701263428}\n",
      "Losses {'ner': 218.72881668806076}\n",
      "Losses {'ner': 217.12707448005676}\n",
      "Losses {'ner': 211.0114300251007}\n",
      "Losses {'ner': 226.36729228496552}\n",
      "Losses {'ner': 214.5805243253708}\n",
      "Losses {'ner': 232.22918558120728}\n",
      "Losses {'ner': 234.45106482505798}\n",
      "Losses {'ner': 227.21963453292847}\n",
      "Losses {'ner': 229.0854799747467}\n",
      "Losses {'ner': 222.18420791625977}\n",
      "Losses {'ner': 214.39253449440002}\n",
      "Losses {'ner': 234.82820415496826}\n",
      "Losses {'ner': 213.53586733341217}\n",
      "Losses {'ner': 223.9589946269989}\n",
      "Losses {'ner': 225.67549920082092}\n",
      "Losses {'ner': 223.1720485687256}\n",
      "Losses {'ner': 235.46119570732117}\n",
      "Losses {'ner': 225.60271525382996}\n",
      "Losses {'ner': 217.2024371623993}\n",
      "Losses {'ner': 226.02379965782166}\n",
      "Losses {'ner': 221.61138740181923}\n",
      "Losses {'ner': 224.0628486275673}\n",
      "Losses {'ner': 214.69909417629242}\n",
      "Losses {'ner': 224.76786088943481}\n",
      "Losses {'ner': 241.8973217010498}\n",
      "Losses {'ner': 231.96240842342377}\n",
      "Losses {'ner': 225.6454360485077}\n",
      "Losses {'ner': 207.540611743927}\n",
      "Losses {'ner': 226.72942519187927}\n",
      "Losses {'ner': 229.20285511016846}\n",
      "Losses {'ner': 207.53844547271729}\n",
      "Losses {'ner': 219.7540899515152}\n",
      "Losses {'ner': 222.20512318611145}\n",
      "Losses {'ner': 221.42441868782043}\n",
      "Losses {'ner': 222.4596230983734}\n",
      "Losses {'ner': 226.68928611278534}\n",
      "Losses {'ner': 214.37474656105042}\n",
      "Losses {'ner': 223.18006944656372}\n",
      "Losses {'ner': 212.35788881778717}\n",
      "Losses {'ner': 228.41949343681335}\n",
      "Losses {'ner': 228.73755407333374}\n",
      "Losses {'ner': 223.96637082099915}\n",
      "Losses {'ner': 217.5358672682196}\n",
      "Losses {'ner': 227.52351832389832}\n",
      "Losses {'ner': 222.16054546833038}\n",
      "Losses {'ner': 222.31783056259155}\n",
      "Losses {'ner': 219.11437463760376}\n",
      "Losses {'ner': 206.20025527477264}\n",
      "Losses {'ner': 228.791002035141}\n",
      "Losses {'ner': 216.12680840492249}\n",
      "Losses {'ner': 219.93672633171082}\n",
      "Losses {'ner': 235.23780608177185}\n",
      "Losses {'ner': 227.66398549079895}\n",
      "Losses {'ner': 209.49372321367264}\n",
      "Losses {'ner': 219.76351594924927}\n",
      "Losses {'ner': 218.59403318166733}\n",
      "Losses {'ner': 228.26440119743347}\n",
      "Losses {'ner': 219.29465371370316}\n",
      "Losses {'ner': 219.80296218395233}\n",
      "Losses {'ner': 229.17839193344116}\n",
      "Losses {'ner': 226.4346101284027}\n",
      "Losses {'ner': 212.9582405090332}\n",
      "Losses {'ner': 225.77544248104095}\n",
      "Losses {'ner': 208.3402922153473}\n",
      "Losses {'ner': 218.63005793094635}\n",
      "Losses {'ner': 220.33883953094482}\n",
      "Losses {'ner': 227.67060327529907}\n",
      "Losses {'ner': 221.36414742469788}\n",
      "Losses {'ner': 219.3755350112915}\n",
      "Losses {'ner': 219.15008449554443}\n",
      "Losses {'ner': 218.182847738266}\n",
      "Losses {'ner': 228.34633839130402}\n",
      "Losses {'ner': 218.77161449193954}\n",
      "Losses {'ner': 231.33821034431458}\n",
      "Losses {'ner': 208.57381558418274}\n",
      "Losses {'ner': 217.10737681388855}\n",
      "Losses {'ner': 210.73096108436584}\n",
      "Losses {'ner': 215.75589966773987}\n",
      "Losses {'ner': 229.1407815515995}\n",
      "Losses {'ner': 216.0052351951599}\n",
      "Losses {'ner': 218.25920748710632}\n",
      "Losses {'ner': 222.7523958683014}\n",
      "Losses {'ner': 221.03235471248627}\n",
      "Losses {'ner': 208.1719925403595}\n",
      "Losses {'ner': 222.8619248867035}\n",
      "Losses {'ner': 232.7407078742981}\n",
      "Losses {'ner': 212.56540966033936}\n",
      "Losses {'ner': 222.90484726428986}\n",
      "Losses {'ner': 224.36307048797607}\n",
      "Losses {'ner': 227.27852261066437}\n",
      "Losses {'ner': 217.20953822135925}\n",
      "Losses {'ner': 219.48252189159393}\n",
      "Losses {'ner': 216.18608677387238}\n",
      "Losses {'ner': 223.28102350234985}\n",
      "Losses {'ner': 222.0766944885254}\n",
      "Losses {'ner': 223.14134633541107}\n",
      "Losses {'ner': 220.48301553726196}\n",
      "Losses {'ner': 226.11104959249496}\n",
      "Losses {'ner': 239.1364313364029}\n",
      "Losses {'ner': 218.13450741767883}\n",
      "Losses {'ner': 223.3258489370346}\n",
      "Losses {'ner': 220.31243976950645}\n",
      "Losses {'ner': 207.22261142730713}\n",
      "Losses {'ner': 223.7833672761917}\n",
      "Losses {'ner': 218.93312978744507}\n",
      "Losses {'ner': 213.40150529146194}\n",
      "Losses {'ner': 222.67623698711395}\n",
      "Losses {'ner': 234.56179285049438}\n",
      "Losses {'ner': 235.70173406600952}\n",
      "Losses {'ner': 232.08041942119598}\n",
      "Losses {'ner': 220.47136545181274}\n",
      "Losses {'ner': 225.30943405628204}\n",
      "Losses {'ner': 233.52685725688934}\n",
      "Losses {'ner': 219.80191510915756}\n",
      "Losses {'ner': 229.88880133628845}\n",
      "Losses {'ner': 207.5262120962143}\n",
      "Losses {'ner': 212.17651641368866}\n",
      "Losses {'ner': 217.22560465335846}\n",
      "Losses {'ner': 212.48525965213776}\n",
      "Losses {'ner': 231.8069338798523}\n",
      "Losses {'ner': 224.44791722297668}\n",
      "Losses {'ner': 227.87749350070953}\n",
      "Losses {'ner': 232.39703941345215}\n",
      "Losses {'ner': 229.62226355075836}\n",
      "Losses {'ner': 209.75517642498016}\n",
      "Losses {'ner': 218.35593438148499}\n",
      "Losses {'ner': 224.5262041091919}\n",
      "Losses {'ner': 232.28142082691193}\n",
      "Losses {'ner': 216.26743245124817}\n",
      "Losses {'ner': 221.62643682956696}\n",
      "Losses {'ner': 219.36080694198608}\n",
      "Losses {'ner': 221.87009990215302}\n",
      "Losses {'ner': 215.5122790336609}\n",
      "Losses {'ner': 211.37868857383728}\n",
      "Losses {'ner': 211.00491881370544}\n",
      "Losses {'ner': 207.38689398765564}\n",
      "Losses {'ner': 223.55912959575653}\n",
      "Losses {'ner': 206.4800145626068}\n",
      "Losses {'ner': 216.4121708869934}\n",
      "Losses {'ner': 220.9242262840271}\n",
      "Losses {'ner': 219.58240979909897}\n",
      "Losses {'ner': 221.15464627742767}\n",
      "Losses {'ner': 225.85867023468018}\n",
      "Losses {'ner': 232.1388566493988}\n",
      "Losses {'ner': 221.8739264011383}\n",
      "Losses {'ner': 221.49924325942993}\n",
      "Losses {'ner': 219.10738575458527}\n",
      "Losses {'ner': 213.3502893447876}\n",
      "Losses {'ner': 193.91394340991974}\n",
      "Losses {'ner': 216.3048164844513}\n",
      "Losses {'ner': 219.6792587041855}\n",
      "Losses {'ner': 217.49644804000854}\n",
      "Losses {'ner': 226.09861302375793}\n",
      "Losses {'ner': 221.32188081741333}\n",
      "Losses {'ner': 223.29227471351624}\n",
      "Losses {'ner': 219.93823546171188}\n",
      "Losses {'ner': 219.62370562553406}\n",
      "Losses {'ner': 221.40586191415787}\n",
      "Losses {'ner': 205.54494547843933}\n",
      "Losses {'ner': 218.52519607543945}\n",
      "Losses {'ner': 214.29195046424866}\n",
      "Losses {'ner': 213.33765816688538}\n",
      "Losses {'ner': 212.54665958881378}\n",
      "Losses {'ner': 203.79535794258118}\n",
      "Losses {'ner': 220.99598294496536}\n",
      "Losses {'ner': 220.63796257972717}\n",
      "Losses {'ner': 217.4598777294159}\n",
      "Losses {'ner': 219.0235537290573}\n",
      "Losses {'ner': 205.88206362724304}\n",
      "Losses {'ner': 209.9465799331665}\n",
      "Losses {'ner': 213.7034500837326}\n",
      "Losses {'ner': 219.67480957508087}\n",
      "Losses {'ner': 215.91896998882294}\n",
      "Losses {'ner': 206.31193816661835}\n",
      "Losses {'ner': 213.1551866531372}\n",
      "Losses {'ner': 215.97826743125916}\n",
      "Losses {'ner': 211.91281628608704}\n",
      "Losses {'ner': 205.24792385101318}\n",
      "Losses {'ner': 221.56491887569427}\n",
      "Losses {'ner': 225.50049078464508}\n",
      "Losses {'ner': 220.54696416854858}\n",
      "Losses {'ner': 231.42723727226257}\n",
      "Losses {'ner': 203.52951765060425}\n",
      "Losses {'ner': 208.53287172317505}\n",
      "Losses {'ner': 221.54878211021423}\n",
      "Losses {'ner': 215.44577825069427}\n",
      "Losses {'ner': 211.0027735233307}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 213.15160429477692}\n",
      "Losses {'ner': 208.40592527389526}\n",
      "Losses {'ner': 217.80078983306885}\n",
      "Losses {'ner': 227.76036036014557}\n",
      "Losses {'ner': 213.87484228610992}\n",
      "Losses {'ner': 209.03973007202148}\n",
      "Losses {'ner': 221.71893501281738}\n",
      "Losses {'ner': 224.4686930179596}\n",
      "Losses {'ner': 226.39426851272583}\n",
      "Losses {'ner': 224.4907832145691}\n",
      "Losses {'ner': 223.9679110646248}\n",
      "Losses {'ner': 216.73783481121063}\n",
      "Losses {'ner': 210.84455585479736}\n",
      "Losses {'ner': 222.09330415725708}\n",
      "Losses {'ner': 221.93060994148254}\n",
      "Losses {'ner': 224.0619996190071}\n",
      "Losses {'ner': 227.7602458000183}\n",
      "Losses {'ner': 220.957581281662}\n",
      "Losses {'ner': 218.98671997711062}\n",
      "Losses {'ner': 226.0086977481842}\n",
      "Losses {'ner': 217.59185314178467}\n",
      "Losses {'ner': 225.9791728258133}\n",
      "Losses {'ner': 222.75218522548676}\n",
      "Losses {'ner': 228.2790004014969}\n",
      "Losses {'ner': 215.51259636878967}\n",
      "Losses {'ner': 206.859089076519}\n",
      "Losses {'ner': 227.99434238672256}\n",
      "Losses {'ner': 218.18142130970955}\n",
      "Losses {'ner': 207.7876694202423}\n",
      "Losses {'ner': 227.6515598297119}\n",
      "Losses {'ner': 210.1477962732315}\n",
      "Losses {'ner': 215.37436878681183}\n",
      "Losses {'ner': 227.5220718383789}\n",
      "Losses {'ner': 215.08117008209229}\n",
      "Losses {'ner': 213.64225041866302}\n",
      "Losses {'ner': 219.07313203811646}\n",
      "Losses {'ner': 225.40202403068542}\n",
      "Losses {'ner': 209.3659518957138}\n",
      "Losses {'ner': 224.13811075687408}\n",
      "Losses {'ner': 209.54469454288483}\n",
      "Losses {'ner': 228.1537947654724}\n",
      "Losses {'ner': 227.0089726448059}\n",
      "Losses {'ner': 215.06755900382996}\n",
      "Losses {'ner': 215.37607645988464}\n",
      "Losses {'ner': 217.12411487102509}\n",
      "Losses {'ner': 220.17877578735352}\n",
      "Losses {'ner': 224.51956868171692}\n",
      "Losses {'ner': 220.70097827911377}\n",
      "Losses {'ner': 227.41042113304138}\n",
      "Losses {'ner': 226.4729244709015}\n",
      "Losses {'ner': 217.7688126564026}\n",
      "Losses {'ner': 215.74173283576965}\n",
      "Losses {'ner': 214.11318945884705}\n",
      "Losses {'ner': 225.49731278419495}\n",
      "Losses {'ner': 224.640726685524}\n",
      "Losses {'ner': 216.59375548362732}\n",
      "Losses {'ner': 225.26417124271393}\n",
      "Losses {'ner': 216.78325295448303}\n",
      "Losses {'ner': 219.60040545463562}\n",
      "Losses {'ner': 203.75246405601501}\n",
      "Losses {'ner': 223.22985863685608}\n",
      "Losses {'ner': 222.26814317703247}\n",
      "Losses {'ner': 219.6453716158867}\n",
      "Losses {'ner': 226.98226952552795}\n",
      "Losses {'ner': 218.8434739112854}\n",
      "Losses {'ner': 215.80392956733704}\n",
      "Losses {'ner': 224.20668995380402}\n",
      "Losses {'ner': 211.74852693080902}\n",
      "Losses {'ner': 222.2950302362442}\n",
      "Losses {'ner': 221.4394520521164}\n",
      "Losses {'ner': 211.36179608106613}\n",
      "Losses {'ner': 238.65760970115662}\n",
      "Losses {'ner': 211.21871161460876}\n",
      "Losses {'ner': 230.6393587589264}\n",
      "Losses {'ner': 206.27597165107727}\n",
      "Losses {'ner': 222.47045278549194}\n",
      "Losses {'ner': 227.46907353401184}\n",
      "Losses {'ner': 217.65106880664825}\n",
      "Losses {'ner': 213.364422082901}\n",
      "Losses {'ner': 225.0311917066574}\n",
      "Losses {'ner': 217.26963794231415}\n",
      "Losses {'ner': 230.17272651195526}\n",
      "Losses {'ner': 219.75452637672424}\n",
      "Losses {'ner': 227.65151166915894}\n",
      "Losses {'ner': 214.0179141163826}\n",
      "Losses {'ner': 224.77785110473633}\n",
      "Losses {'ner': 207.65572881698608}\n",
      "Losses {'ner': 210.06609177589417}\n",
      "Losses {'ner': 227.6959799528122}\n",
      "Losses {'ner': 224.21787786483765}\n",
      "Losses {'ner': 218.76941323280334}\n",
      "Losses {'ner': 212.7534520626068}\n",
      "Losses {'ner': 228.40523266792297}\n",
      "Losses {'ner': 202.33446741104126}\n",
      "Losses {'ner': 210.9989402294159}\n",
      "Losses {'ner': 215.62747514247894}\n",
      "Losses {'ner': 223.16895651817322}\n",
      "Losses {'ner': 218.7220344543457}\n",
      "Losses {'ner': 226.97042679786682}\n",
      "Losses {'ner': 220.07306843996048}\n",
      "Losses {'ner': 200.497012257576}\n",
      "Losses {'ner': 211.2222352027893}\n",
      "Losses {'ner': 213.27374631166458}\n",
      "Losses {'ner': 218.36416363716125}\n",
      "Losses {'ner': 225.45720911026}\n",
      "Losses {'ner': 234.6703370809555}\n",
      "Losses {'ner': 226.85034310817719}\n",
      "Losses {'ner': 221.10309767723083}\n",
      "Losses {'ner': 238.24471592903137}\n",
      "Losses {'ner': 221.1219766139984}\n",
      "Losses {'ner': 230.70424914360046}\n",
      "Losses {'ner': 219.44361460208893}\n",
      "Losses {'ner': 225.92534792423248}\n",
      "Losses {'ner': 232.17575764656067}\n",
      "Losses {'ner': 230.0255880355835}\n",
      "Losses {'ner': 221.13498532772064}\n",
      "Losses {'ner': 211.84692201018333}\n",
      "Losses {'ner': 212.2843952178955}\n",
      "Losses {'ner': 218.39153110980988}\n",
      "Losses {'ner': 224.38859820365906}\n",
      "Losses {'ner': 215.69765448570251}\n",
      "Losses {'ner': 221.36633777618408}\n",
      "Losses {'ner': 208.19708859920502}\n",
      "Losses {'ner': 203.8255935907364}\n",
      "Losses {'ner': 200.2350596189499}\n",
      "Losses {'ner': 216.53129971027374}\n",
      "Losses {'ner': 226.89409518241882}\n",
      "Losses {'ner': 221.70606112480164}\n",
      "Losses {'ner': 227.49854254722595}\n",
      "Losses {'ner': 224.90032297372818}\n",
      "Losses {'ner': 232.79395627975464}\n",
      "Losses {'ner': 212.10748463869095}\n",
      "Losses {'ner': 221.36077892780304}\n",
      "Losses {'ner': 221.02581536769867}\n",
      "Losses {'ner': 222.9777479171753}\n",
      "Losses {'ner': 211.36371755599976}\n",
      "Losses {'ner': 220.64113545417786}\n",
      "Losses {'ner': 209.9080090522766}\n",
      "Losses {'ner': 221.9045250415802}\n",
      "Losses {'ner': 219.07307183742523}\n",
      "Losses {'ner': 206.8820824623108}\n",
      "Losses {'ner': 207.98085963726044}\n",
      "Losses {'ner': 216.48500442504883}\n",
      "Losses {'ner': 200.65707314014435}\n",
      "Losses {'ner': 199.33428871631622}\n",
      "Losses {'ner': 219.54369950294495}\n",
      "Losses {'ner': 216.77227234840393}\n",
      "Losses {'ner': 205.26744502782822}\n",
      "Losses {'ner': 226.13690328598022}\n",
      "Losses {'ner': 224.20999360084534}\n",
      "Losses {'ner': 226.08760452270508}\n",
      "Losses {'ner': 207.6445963382721}\n",
      "Losses {'ner': 218.16829800605774}\n",
      "Losses {'ner': 229.12186205387115}\n",
      "Losses {'ner': 229.81568901985884}\n",
      "Losses {'ner': 221.3316056728363}\n",
      "Losses {'ner': 222.0499017238617}\n",
      "Losses {'ner': 225.0648832321167}\n",
      "Losses {'ner': 219.11353981494904}\n",
      "Losses {'ner': 224.02049803733826}\n",
      "Losses {'ner': 226.7189302444458}\n",
      "Losses {'ner': 224.8276879787445}\n",
      "Losses {'ner': 217.62284231185913}\n",
      "Losses {'ner': 222.73914754390717}\n",
      "Losses {'ner': 212.67330253124237}\n",
      "Losses {'ner': 218.29773807525635}\n",
      "Losses {'ner': 207.04193210601807}\n",
      "Losses {'ner': 228.34764218330383}\n",
      "Losses {'ner': 223.0335333943367}\n",
      "Losses {'ner': 227.31679677963257}\n",
      "Losses {'ner': 222.42929363250732}\n",
      "Losses {'ner': 230.81222414970398}\n",
      "Losses {'ner': 223.5607430934906}\n",
      "Losses {'ner': 233.0408935546875}\n",
      "Losses {'ner': 210.80722963809967}\n",
      "Losses {'ner': 230.86721539497375}\n",
      "Losses {'ner': 221.94999170303345}\n",
      "Losses {'ner': 228.1648508310318}\n",
      "Losses {'ner': 217.5649163722992}\n",
      "Losses {'ner': 225.0114048719406}\n",
      "Losses {'ner': 219.2747733592987}\n",
      "Losses {'ner': 220.82524371147156}\n",
      "Losses {'ner': 217.67338061332703}\n",
      "Losses {'ner': 216.28719639778137}\n",
      "Losses {'ner': 222.22264122962952}\n",
      "Losses {'ner': 222.9857348203659}\n",
      "Losses {'ner': 214.21822559833527}\n",
      "Losses {'ner': 204.12127125263214}\n",
      "Losses {'ner': 219.19947504997253}\n",
      "Losses {'ner': 222.61141729354858}\n",
      "Losses {'ner': 221.84577584266663}\n",
      "Losses {'ner': 220.5088233947754}\n",
      "Losses {'ner': 217.76180112361908}\n",
      "Losses {'ner': 220.03111362457275}\n",
      "Losses {'ner': 221.63234043121338}\n",
      "Losses {'ner': 216.28030014038086}\n",
      "Losses {'ner': 227.7466208934784}\n",
      "Losses {'ner': 218.05392241477966}\n",
      "Losses {'ner': 232.36580657958984}\n",
      "Losses {'ner': 213.29148656129837}\n",
      "Losses {'ner': 215.09278118610382}\n",
      "Losses {'ner': 230.67283082008362}\n",
      "Losses {'ner': 222.19296503067017}\n",
      "Losses {'ner': 224.45472586154938}\n",
      "Losses {'ner': 206.16857385635376}\n",
      "Losses {'ner': 219.92648124694824}\n",
      "Losses {'ner': 212.20816779136658}\n",
      "Losses {'ner': 231.6302216053009}\n",
      "Losses {'ner': 233.08195281028748}\n",
      "Losses {'ner': 226.7272496521473}\n",
      "Losses {'ner': 225.08692064881325}\n",
      "Losses {'ner': 222.38552522659302}\n",
      "Losses {'ner': 219.30265176296234}\n",
      "Losses {'ner': 213.01528191566467}\n",
      "Losses {'ner': 218.37217009067535}\n",
      "Losses {'ner': 228.9066776446998}\n",
      "Losses {'ner': 207.5180402994156}\n",
      "Losses {'ner': 223.96302223205566}\n",
      "Losses {'ner': 223.317791223526}\n",
      "Losses {'ner': 227.72580432891846}\n",
      "Losses {'ner': 211.99752354621887}\n",
      "Losses {'ner': 231.40422189235687}\n",
      "Losses {'ner': 218.05865740776062}\n",
      "Losses {'ner': 220.58899354934692}\n",
      "Losses {'ner': 209.30108320713043}\n",
      "Losses {'ner': 221.8670448064804}\n",
      "Losses {'ner': 231.76780152320862}\n",
      "Losses {'ner': 208.7089251279831}\n",
      "Losses {'ner': 215.62051677703857}\n",
      "Losses {'ner': 228.9385632276535}\n",
      "Losses {'ner': 226.0560109615326}\n",
      "Losses {'ner': 218.841055393219}\n",
      "Losses {'ner': 210.16553962230682}\n",
      "Losses {'ner': 220.7451615333557}\n",
      "Losses {'ner': 219.12505173683167}\n",
      "Losses {'ner': 212.0800597667694}\n",
      "Losses {'ner': 211.1793406009674}\n",
      "Losses {'ner': 235.12139678001404}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 218.5905406177044}\n",
      "Losses {'ner': 227.62916457653046}\n",
      "Losses {'ner': 228.91239178180695}\n",
      "Losses {'ner': 220.10667157173157}\n",
      "Losses {'ner': 226.57473158836365}\n",
      "Losses {'ner': 217.38730812072754}\n",
      "Losses {'ner': 230.94501614570618}\n",
      "Losses {'ner': 211.88739037513733}\n",
      "Losses {'ner': 223.97543025016785}\n",
      "Losses {'ner': 217.13701057434082}\n",
      "Losses {'ner': 223.44409811496735}\n",
      "Losses {'ner': 230.58414483070374}\n",
      "Losses {'ner': 219.73332154750824}\n",
      "Losses {'ner': 236.73600912094116}\n",
      "Losses {'ner': 222.48322415351868}\n",
      "Losses {'ner': 220.29402613639832}\n",
      "Losses {'ner': 218.93549275398254}\n",
      "Losses {'ner': 223.9080754518509}\n",
      "Losses {'ner': 201.12749165296555}\n",
      "Losses {'ner': 212.316526055336}\n",
      "Losses {'ner': 210.6017985343933}\n",
      "Losses {'ner': 216.11473262310028}\n",
      "Losses {'ner': 226.27823567390442}\n",
      "Losses {'ner': 231.97292685508728}\n",
      "Losses {'ner': 230.25530767440796}\n",
      "Losses {'ner': 222.46332216262817}\n",
      "Losses {'ner': 227.5088073015213}\n",
      "Losses {'ner': 214.84898507595062}\n",
      "Losses {'ner': 210.95838969945908}\n",
      "Losses {'ner': 226.90749764442444}\n",
      "Losses {'ner': 220.69530647993088}\n",
      "Losses {'ner': 224.07728517055511}\n",
      "Losses {'ner': 216.17617046833038}\n",
      "Losses {'ner': 217.26590979099274}\n",
      "Losses {'ner': 218.61694991588593}\n",
      "Losses {'ner': 222.1136989593506}\n",
      "Losses {'ner': 211.55603551864624}\n",
      "Losses {'ner': 210.9407558441162}\n",
      "Losses {'ner': 219.87013864517212}\n",
      "Losses {'ner': 213.66246655583382}\n",
      "Losses {'ner': 221.76371276378632}\n",
      "Losses {'ner': 207.6435307264328}\n",
      "Losses {'ner': 230.7253565788269}\n",
      "Losses {'ner': 204.7908217906952}\n",
      "Losses {'ner': 220.59830236434937}\n",
      "Losses {'ner': 222.28666949272156}\n",
      "Losses {'ner': 219.9694126546383}\n",
      "Losses {'ner': 215.06960463523865}\n",
      "Losses {'ner': 217.58334624767303}\n",
      "Losses {'ner': 218.4517204761505}\n",
      "Losses {'ner': 216.5949187874794}\n",
      "Losses {'ner': 220.43304240703583}\n",
      "Losses {'ner': 215.12844371795654}\n",
      "Losses {'ner': 221.62855124473572}\n",
      "Losses {'ner': 209.63269543647766}\n",
      "Losses {'ner': 233.9438681602478}\n",
      "Losses {'ner': 211.40621531009674}\n",
      "Losses {'ner': 215.98071384429932}\n",
      "Losses {'ner': 215.9175102710724}\n",
      "Losses {'ner': 208.57782292366028}\n",
      "Losses {'ner': 222.5005340576172}\n",
      "Losses {'ner': 222.20261943340302}\n",
      "Losses {'ner': 236.7488603591919}\n",
      "Losses {'ner': 208.4561403989792}\n",
      "Losses {'ner': 224.91189879179}\n",
      "Losses {'ner': 227.28427064418793}\n",
      "Losses {'ner': 214.61476874351501}\n",
      "Losses {'ner': 218.7471661567688}\n",
      "Losses {'ner': 220.1772519350052}\n",
      "Losses {'ner': 224.6450846195221}\n",
      "Losses {'ner': 213.45107281208038}\n",
      "Losses {'ner': 225.7082417011261}\n",
      "Losses {'ner': 215.85038328170776}\n",
      "Losses {'ner': 212.05029892921448}\n",
      "Losses {'ner': 228.48727989196777}\n",
      "Losses {'ner': 226.37689459323883}\n",
      "Losses {'ner': 223.54200434684753}\n",
      "Losses {'ner': 220.84198105335236}\n",
      "Losses {'ner': 221.53560522943735}\n",
      "Losses {'ner': 220.3713023662567}\n",
      "Losses {'ner': 216.15930199623108}\n",
      "Losses {'ner': 212.505267649889}\n",
      "Losses {'ner': 213.90939497947693}\n",
      "Losses {'ner': 216.05479335784912}\n",
      "Losses {'ner': 226.24445700645447}\n",
      "Losses {'ner': 227.47803580760956}\n",
      "Losses {'ner': 211.8393518924713}\n",
      "Losses {'ner': 220.84689116477966}\n",
      "Losses {'ner': 223.24078142642975}\n",
      "Losses {'ner': 210.1608099937439}\n",
      "Losses {'ner': 223.1371259689331}\n",
      "Losses {'ner': 217.82431960105896}\n",
      "Losses {'ner': 208.39790964126587}\n",
      "Losses {'ner': 239.74916529655457}\n",
      "Losses {'ner': 231.35660231113434}\n",
      "Losses {'ner': 219.27979934215546}\n",
      "Losses {'ner': 215.73514807224274}\n",
      "Losses {'ner': 218.16826832294464}\n",
      "Losses {'ner': 205.53557324409485}\n",
      "Losses {'ner': 223.15584123134613}\n",
      "Losses {'ner': 205.4480528831482}\n",
      "Losses {'ner': 220.6259400844574}\n",
      "Losses {'ner': 230.18743252754211}\n",
      "Losses {'ner': 221.85620856285095}\n",
      "Losses {'ner': 222.26251304149628}\n",
      "Losses {'ner': 218.28515088558197}\n",
      "Losses {'ner': 227.37042379379272}\n",
      "Losses {'ner': 214.04986310005188}\n",
      "Losses {'ner': 222.61722946166992}\n",
      "Losses {'ner': 227.97280383110046}\n",
      "Losses {'ner': 231.36529898643494}\n",
      "Losses {'ner': 229.4879161119461}\n",
      "Losses {'ner': 215.8175404071808}\n",
      "Losses {'ner': 209.35864341259003}\n",
      "Losses {'ner': 227.7845377922058}\n",
      "Losses {'ner': 217.01787114143372}\n",
      "Losses {'ner': 210.76865792274475}\n",
      "Losses {'ner': 221.13830518722534}\n",
      "Losses {'ner': 231.62017959356308}\n",
      "Losses {'ner': 212.02249026298523}\n",
      "Losses {'ner': 217.63935804367065}\n",
      "Losses {'ner': 216.2076051235199}\n",
      "Losses {'ner': 225.32897520065308}\n",
      "Losses {'ner': 219.67981505393982}\n",
      "Losses {'ner': 212.53588342666626}\n",
      "Losses {'ner': 229.0993251800537}\n",
      "Losses {'ner': 226.80315899848938}\n",
      "Losses {'ner': 230.0090618133545}\n",
      "Losses {'ner': 215.0049889087677}\n",
      "Losses {'ner': 227.56048369407654}\n",
      "Losses {'ner': 223.09362483024597}\n",
      "Losses {'ner': 217.85509932041168}\n",
      "Losses {'ner': 223.72036957740784}\n",
      "Losses {'ner': 223.71453523635864}\n",
      "Losses {'ner': 222.8961808681488}\n",
      "Losses {'ner': 209.0607488155365}\n",
      "Losses {'ner': 218.91622585058212}\n",
      "Losses {'ner': 213.94549596309662}\n",
      "Losses {'ner': 206.62891793251038}\n",
      "Losses {'ner': 211.43304419517517}\n",
      "Losses {'ner': 227.01964962482452}\n",
      "Losses {'ner': 217.2590096592903}\n",
      "Losses {'ner': 224.28967916965485}\n",
      "Losses {'ner': 223.92899352312088}\n",
      "Losses {'ner': 217.7833297252655}\n",
      "Losses {'ner': 220.92101514339447}\n",
      "Losses {'ner': 214.95651423931122}\n",
      "Losses {'ner': 217.84264469146729}\n",
      "Losses {'ner': 217.19472217559814}\n",
      "Losses {'ner': 229.8302721977234}\n",
      "Losses {'ner': 210.91193717718124}\n",
      "Losses {'ner': 215.6477505862713}\n",
      "Losses {'ner': 221.2746358513832}\n",
      "Losses {'ner': 222.4703266620636}\n",
      "Losses {'ner': 217.8412356376648}\n",
      "Losses {'ner': 213.10941624641418}\n",
      "Losses {'ner': 222.9019936323166}\n",
      "Losses {'ner': 215.49195456504822}\n",
      "Losses {'ner': 207.0572544336319}\n",
      "Losses {'ner': 218.40948975086212}\n",
      "Losses {'ner': 233.37433457374573}\n",
      "Losses {'ner': 215.82313108444214}\n",
      "Losses {'ner': 225.96439123153687}\n",
      "Losses {'ner': 215.02916479110718}\n",
      "Losses {'ner': 219.48091840744019}\n",
      "Losses {'ner': 218.68956685066223}\n",
      "Losses {'ner': 220.22918176651}\n",
      "Losses {'ner': 220.4561710357666}\n",
      "Losses {'ner': 213.93966495990753}\n",
      "Losses {'ner': 220.18993949890137}\n",
      "Losses {'ner': 225.10960614681244}\n",
      "Losses {'ner': 220.71009039878845}\n",
      "Losses {'ner': 213.7325245141983}\n",
      "Losses {'ner': 221.9716022014618}\n",
      "Losses {'ner': 234.08053481578827}\n",
      "Losses {'ner': 226.62960720062256}\n",
      "Losses {'ner': 221.22188258171082}\n",
      "Losses {'ner': 218.67697596549988}\n",
      "Losses {'ner': 233.62112438678741}\n",
      "Losses {'ner': 217.28962421417236}\n",
      "Losses {'ner': 218.11300611495972}\n",
      "Losses {'ner': 234.37420320510864}\n",
      "Losses {'ner': 220.28280019760132}\n",
      "Losses {'ner': 220.2999234199524}\n",
      "Losses {'ner': 221.01116597652435}\n",
      "Losses {'ner': 217.39170336723328}\n",
      "Losses {'ner': 212.46490973234177}\n",
      "Losses {'ner': 204.9532699584961}\n",
      "Losses {'ner': 210.00984466075897}\n",
      "Losses {'ner': 228.882142663002}\n",
      "Losses {'ner': 214.57126450538635}\n",
      "Losses {'ner': 226.28862896561623}\n",
      "Losses {'ner': 216.12278497219086}\n",
      "Losses {'ner': 214.15667688846588}\n",
      "Losses {'ner': 217.34943288564682}\n",
      "Losses {'ner': 224.54247164726257}\n",
      "Losses {'ner': 226.8792177438736}\n",
      "Losses {'ner': 231.74199032783508}\n",
      "Losses {'ner': 209.1733775138855}\n",
      "Losses {'ner': 227.33087468147278}\n",
      "Losses {'ner': 224.65530651807785}\n",
      "Losses {'ner': 214.50548219680786}\n",
      "Losses {'ner': 209.594708442688}\n",
      "Losses {'ner': 215.08736276626587}\n",
      "Losses {'ner': 197.90323999524117}\n",
      "Losses {'ner': 219.47091209888458}\n",
      "Losses {'ner': 218.70697104930878}\n",
      "Losses {'ner': 228.8512214422226}\n",
      "Losses {'ner': 215.46412670612335}\n",
      "Losses {'ner': 229.29440236091614}\n",
      "Losses {'ner': 227.18235445022583}\n",
      "Losses {'ner': 215.79702591896057}\n",
      "Losses {'ner': 221.3940371274948}\n",
      "Losses {'ner': 226.3441767692566}\n",
      "Losses {'ner': 228.81273084878922}\n",
      "Losses {'ner': 225.901824593544}\n",
      "Losses {'ner': 198.70497298240662}\n",
      "Losses {'ner': 226.881982088089}\n",
      "Losses {'ner': 221.93302804231644}\n",
      "Losses {'ner': 206.92352986335754}\n",
      "Losses {'ner': 226.09966397285461}\n",
      "Losses {'ner': 217.52012360095978}\n",
      "Losses {'ner': 210.89745831489563}\n",
      "Losses {'ner': 218.26416552066803}\n",
      "Losses {'ner': 232.3612766265869}\n",
      "Losses {'ner': 230.16278660297394}\n",
      "Losses {'ner': 224.92388200759888}\n",
      "Losses {'ner': 215.85300993919373}\n",
      "Losses {'ner': 218.62361347675323}\n",
      "Losses {'ner': 217.6011519432068}\n",
      "Losses {'ner': 213.3720407485962}\n",
      "Losses {'ner': 209.2233110666275}\n",
      "Losses {'ner': 227.2552514076233}\n",
      "Losses {'ner': 223.76352846622467}\n",
      "Losses {'ner': 219.3991950750351}\n",
      "Losses {'ner': 217.45935010910034}\n",
      "Losses {'ner': 224.8013994693756}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 220.62972593307495}\n",
      "Losses {'ner': 222.8529440164566}\n",
      "Losses {'ner': 197.84688940644264}\n",
      "Losses {'ner': 218.26342749595642}\n",
      "Losses {'ner': 211.24114310741425}\n",
      "Losses {'ner': 215.2584331035614}\n",
      "Losses {'ner': 229.55855298042297}\n",
      "Losses {'ner': 222.93589758872986}\n",
      "Losses {'ner': 212.90589118003845}\n",
      "Losses {'ner': 218.71376419067383}\n",
      "Losses {'ner': 208.27241492271423}\n",
      "Losses {'ner': 216.14632415771484}\n",
      "Losses {'ner': 225.28029441833496}\n",
      "Losses {'ner': 209.42730069160461}\n",
      "Losses {'ner': 221.15283769369125}\n",
      "Losses {'ner': 216.20725524425507}\n",
      "Losses {'ner': 214.68125236034393}\n",
      "Losses {'ner': 218.98608100414276}\n",
      "Losses {'ner': 212.90968942642212}\n",
      "Losses {'ner': 229.37630605697632}\n",
      "Losses {'ner': 211.58666515350342}\n",
      "Losses {'ner': 215.33537101745605}\n",
      "Losses {'ner': 229.29131335020065}\n",
      "Losses {'ner': 208.6782791018486}\n",
      "Losses {'ner': 214.455650806427}\n",
      "Losses {'ner': 220.66526287794113}\n",
      "Losses {'ner': 223.09471154212952}\n",
      "Losses {'ner': 219.47706151008606}\n",
      "Losses {'ner': 234.6564269065857}\n",
      "Losses {'ner': 204.17389911413193}\n",
      "Losses {'ner': 206.6247524023056}\n",
      "Losses {'ner': 224.70998001098633}\n",
      "Losses {'ner': 211.65726280212402}\n",
      "Losses {'ner': 200.72847259044647}\n",
      "Losses {'ner': 221.1071116924286}\n",
      "Losses {'ner': 207.13430726528168}\n",
      "Losses {'ner': 212.27741199731827}\n",
      "Losses {'ner': 216.12511950731277}\n",
      "Losses {'ner': 232.74770641326904}\n",
      "Losses {'ner': 225.97683346271515}\n",
      "Losses {'ner': 210.29040670394897}\n",
      "Losses {'ner': 205.61678525805473}\n",
      "Losses {'ner': 210.47408533096313}\n",
      "Losses {'ner': 224.075146317482}\n",
      "Losses {'ner': 208.91970229148865}\n",
      "Losses {'ner': 223.9130156636238}\n",
      "Losses {'ner': 218.2684034705162}\n",
      "Losses {'ner': 223.61022996902466}\n",
      "Losses {'ner': 217.87588024139404}\n",
      "Losses {'ner': 220.66957068443298}\n",
      "Losses {'ner': 212.0691750049591}\n",
      "Losses {'ner': 212.7582551240921}\n",
      "Losses {'ner': 207.26866847276688}\n",
      "Losses {'ner': 223.18901419639587}\n",
      "Losses {'ner': 222.8601381778717}\n",
      "Losses {'ner': 211.00596499443054}\n",
      "Losses {'ner': 214.66040325164795}\n",
      "Losses {'ner': 231.40259164571762}\n",
      "Losses {'ner': 220.21939396858215}\n",
      "Losses {'ner': 229.93873810768127}\n",
      "Losses {'ner': 222.86145949363708}\n",
      "Losses {'ner': 228.63929891586304}\n",
      "Losses {'ner': 217.91364169120789}\n",
      "Losses {'ner': 209.53707551956177}\n",
      "Losses {'ner': 214.9759225845337}\n",
      "Losses {'ner': 226.51659959554672}\n",
      "Losses {'ner': 199.46762484312057}\n",
      "Losses {'ner': 221.500385761261}\n",
      "Losses {'ner': 220.52029073238373}\n",
      "Losses {'ner': 221.79822325706482}\n",
      "Losses {'ner': 218.654629945755}\n",
      "Losses {'ner': 217.0422923564911}\n",
      "Losses {'ner': 224.2794172167778}\n",
      "Losses {'ner': 227.09828388690948}\n",
      "Losses {'ner': 217.21812641620636}\n",
      "Losses {'ner': 226.73957657814026}\n",
      "Losses {'ner': 217.35675486922264}\n",
      "Losses {'ner': 222.76322984695435}\n",
      "Losses {'ner': 202.68430280685425}\n",
      "Losses {'ner': 230.75990444421768}\n",
      "Losses {'ner': 214.1860280930996}\n",
      "Losses {'ner': 223.59710916876793}\n",
      "Losses {'ner': 222.83993554115295}\n",
      "Losses {'ner': 222.46575075387955}\n",
      "Losses {'ner': 209.00914001464844}\n",
      "Losses {'ner': 215.44051957130432}\n",
      "Losses {'ner': 216.27956128120422}\n",
      "Losses {'ner': 209.8242028951645}\n",
      "Losses {'ner': 222.63623797893524}\n",
      "Losses {'ner': 215.2633455991745}\n",
      "Losses {'ner': 209.03627157211304}\n",
      "Losses {'ner': 215.024596452713}\n",
      "Losses {'ner': 224.43542790412903}\n",
      "Losses {'ner': 218.80031156539917}\n",
      "Losses {'ner': 216.92823433876038}\n",
      "Losses {'ner': 191.03712943196297}\n",
      "Losses {'ner': 226.16343212127686}\n",
      "Losses {'ner': 217.29545783996582}\n",
      "Losses {'ner': 224.08916664123535}\n",
      "Losses {'ner': 230.53902459144592}\n",
      "Losses {'ner': 222.66484308242798}\n",
      "Losses {'ner': 218.71482348442078}\n",
      "Losses {'ner': 228.73106503486633}\n",
      "Losses {'ner': 216.73565071821213}\n",
      "Losses {'ner': 232.89863109588623}\n",
      "Losses {'ner': 228.83149588108063}\n",
      "Losses {'ner': 217.60099697113037}\n",
      "Losses {'ner': 226.37488859891891}\n",
      "Losses {'ner': 212.2092917561531}\n",
      "Losses {'ner': 229.01126039028168}\n",
      "Losses {'ner': 224.6468186378479}\n",
      "Losses {'ner': 235.12239742279053}\n",
      "Losses {'ner': 215.20697569847107}\n",
      "Losses {'ner': 222.96329855918884}\n",
      "Losses {'ner': 215.15627312660217}\n",
      "Losses {'ner': 214.29038071632385}\n",
      "Losses {'ner': 219.47103929519653}\n",
      "Losses {'ner': 221.8419145345688}\n",
      "Losses {'ner': 223.23774361610413}\n",
      "Losses {'ner': 230.25661838054657}\n",
      "Losses {'ner': 226.96451556682587}\n",
      "Losses {'ner': 231.54055881500244}\n",
      "Losses {'ner': 221.31204921007156}\n",
      "Losses {'ner': 228.66856068372726}\n",
      "Losses {'ner': 219.71665382385254}\n",
      "Losses {'ner': 223.5670189857483}\n",
      "Losses {'ner': 210.60217261314392}\n",
      "Losses {'ner': 206.3476767539978}\n",
      "Losses {'ner': 216.36882030963898}\n",
      "Losses {'ner': 234.00519561767578}\n",
      "Losses {'ner': 206.955885887146}\n",
      "Losses {'ner': 223.0721994638443}\n",
      "Losses {'ner': 217.44534826278687}\n",
      "Losses {'ner': 215.31340980529785}\n",
      "Losses {'ner': 222.75031054019928}\n",
      "Losses {'ner': 227.84939980506897}\n",
      "Losses {'ner': 230.22204315662384}\n",
      "Losses {'ner': 216.5715879201889}\n",
      "Losses {'ner': 220.45810735225677}\n",
      "Losses {'ner': 213.00893926620483}\n",
      "Losses {'ner': 220.1117753982544}\n",
      "Losses {'ner': 219.9687567949295}\n",
      "Losses {'ner': 232.5616750717163}\n",
      "Losses {'ner': 223.285942196846}\n",
      "Losses {'ner': 220.97286415100098}\n",
      "Losses {'ner': 224.91285622119904}\n",
      "Losses {'ner': 227.1089961528778}\n",
      "Losses {'ner': 228.07514071464539}\n",
      "Losses {'ner': 221.07863783836365}\n",
      "Losses {'ner': 222.20076155662537}\n",
      "Losses {'ner': 224.77128100395203}\n",
      "Losses {'ner': 227.47018766403198}\n",
      "Losses {'ner': 213.84100317955017}\n",
      "Losses {'ner': 228.30307006835938}\n",
      "Losses {'ner': 228.7201704978943}\n",
      "Losses {'ner': 219.06962943077087}\n",
      "Losses {'ner': 216.7175703048706}\n",
      "Losses {'ner': 226.9048856496811}\n",
      "Losses {'ner': 226.3902314901352}\n",
      "Losses {'ner': 234.08049750328064}\n",
      "Losses {'ner': 216.07096326351166}\n",
      "Losses {'ner': 219.0311050415039}\n",
      "Losses {'ner': 217.52337390184402}\n",
      "Losses {'ner': 204.44399785995483}\n",
      "Losses {'ner': 218.3161816596985}\n",
      "Losses {'ner': 217.859757065773}\n",
      "Losses {'ner': 214.09993731975555}\n",
      "Losses {'ner': 222.98486870527267}\n",
      "Losses {'ner': 222.8261058330536}\n",
      "Losses {'ner': 226.74498558044434}\n",
      "Losses {'ner': 223.66010546684265}\n",
      "Losses {'ner': 213.85823440551758}\n",
      "Losses {'ner': 236.5572853088379}\n",
      "Losses {'ner': 221.7217619419098}\n",
      "Losses {'ner': 225.96753120422363}\n",
      "Losses {'ner': 224.2677607536316}\n",
      "Losses {'ner': 222.81518626213074}\n",
      "Losses {'ner': 223.00706005096436}\n",
      "Losses {'ner': 217.41766130924225}\n",
      "Losses {'ner': 216.52495789527893}\n",
      "Losses {'ner': 227.81951808929443}\n",
      "Losses {'ner': 215.23889619112015}\n",
      "Losses {'ner': 226.1358621120453}\n",
      "Losses {'ner': 212.36005556583405}\n",
      "Losses {'ner': 215.84041512012482}\n",
      "Losses {'ner': 227.74026483297348}\n",
      "Losses {'ner': 220.21318328380585}\n",
      "Losses {'ner': 216.83996498584747}\n",
      "Losses {'ner': 234.44718420505524}\n",
      "Losses {'ner': 215.96279680728912}\n",
      "Losses {'ner': 217.380979180336}\n",
      "Losses {'ner': 221.85693591833115}\n",
      "Losses {'ner': 235.8911361694336}\n",
      "Losses {'ner': 225.633034825325}\n",
      "Losses {'ner': 216.65554130077362}\n",
      "Losses {'ner': 208.5245715379715}\n",
      "Losses {'ner': 222.97468948364258}\n",
      "Losses {'ner': 214.10704731941223}\n",
      "Losses {'ner': 213.58928108215332}\n",
      "Losses {'ner': 234.21883988380432}\n",
      "Losses {'ner': 230.2480486035347}\n",
      "Losses {'ner': 226.5304902791977}\n",
      "Losses {'ner': 220.8203216791153}\n",
      "Losses {'ner': 228.3393280506134}\n",
      "Losses {'ner': 224.93390655517578}\n",
      "Losses {'ner': 216.07485949993134}\n",
      "Losses {'ner': 241.0900912284851}\n",
      "Losses {'ner': 221.60448801517487}\n",
      "Losses {'ner': 226.58598065376282}\n",
      "Losses {'ner': 222.9765648841858}\n",
      "Losses {'ner': 226.72909128665924}\n",
      "Losses {'ner': 217.22544729709625}\n",
      "Losses {'ner': 230.73772859573364}\n",
      "Losses {'ner': 215.01781949400902}\n",
      "Losses {'ner': 225.49462485313416}\n",
      "Losses {'ner': 220.13876321911812}\n",
      "Losses {'ner': 225.27842354774475}\n",
      "Losses {'ner': 225.13052016496658}\n",
      "Losses {'ner': 219.82120740413666}\n",
      "Losses {'ner': 220.5855473279953}\n",
      "Losses {'ner': 222.79364728927612}\n",
      "Losses {'ner': 217.4161981344223}\n",
      "Losses {'ner': 231.38519263267517}\n",
      "Losses {'ner': 234.95174026489258}\n",
      "Losses {'ner': 222.9322316646576}\n",
      "Losses {'ner': 228.7999291419983}\n",
      "Losses {'ner': 213.5731211900711}\n",
      "Losses {'ner': 233.3203318119049}\n",
      "Losses {'ner': 215.7907862663269}\n",
      "Losses {'ner': 225.68941712379456}\n",
      "Losses {'ner': 229.075355052948}\n",
      "Losses {'ner': 223.32975506782532}\n",
      "Losses {'ner': 230.71429878473282}\n",
      "Losses {'ner': 220.89459264278412}\n",
      "Losses {'ner': 232.07612466812134}\n",
      "Losses {'ner': 222.95843887329102}\n",
      "Losses {'ner': 214.58093965053558}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 215.5148766040802}\n",
      "Losses {'ner': 220.74970591068268}\n",
      "Losses {'ner': 208.41503369808197}\n",
      "Losses {'ner': 214.34111869335175}\n",
      "Losses {'ner': 215.6681432723999}\n",
      "Losses {'ner': 213.10866856575012}\n",
      "Losses {'ner': 205.51891446113586}\n",
      "Losses {'ner': 220.11182177066803}\n",
      "Losses {'ner': 220.60057640075684}\n",
      "Losses {'ner': 214.86819744110107}\n",
      "Losses {'ner': 217.99861586093903}\n",
      "Losses {'ner': 221.8057712316513}\n",
      "Losses {'ner': 205.0799639225006}\n",
      "Losses {'ner': 204.1726714372635}\n",
      "Losses {'ner': 213.83682942390442}\n",
      "Losses {'ner': 222.84376913309097}\n",
      "Losses {'ner': 216.09584468603134}\n",
      "Losses {'ner': 218.21271061897278}\n",
      "Losses {'ner': 222.27120846509933}\n",
      "Losses {'ner': 217.12183928489685}\n",
      "Losses {'ner': 222.20482313632965}\n",
      "Losses {'ner': 224.16139233112335}\n",
      "Losses {'ner': 218.2612180709839}\n",
      "Losses {'ner': 231.1634407043457}\n",
      "Losses {'ner': 207.20241248607635}\n",
      "Losses {'ner': 221.23184114694595}\n",
      "Losses {'ner': 220.4111385345459}\n",
      "Losses {'ner': 218.48994183540344}\n",
      "Losses {'ner': 224.94992542266846}\n",
      "Losses {'ner': 222.3546177148819}\n",
      "Losses {'ner': 213.11202192306519}\n",
      "Losses {'ner': 226.6537482738495}\n",
      "Losses {'ner': 211.24477624893188}\n",
      "Losses {'ner': 226.67795383930206}\n",
      "Losses {'ner': 220.2749092578888}\n",
      "Losses {'ner': 221.4985044002533}\n",
      "Losses {'ner': 228.56697130203247}\n",
      "Losses {'ner': 215.43464946746826}\n",
      "Losses {'ner': 218.46243929862976}\n",
      "Losses {'ner': 206.14403092861176}\n",
      "Losses {'ner': 217.1304823756218}\n",
      "Losses {'ner': 219.4858775138855}\n",
      "Losses {'ner': 223.96516931056976}\n",
      "Losses {'ner': 226.75308179855347}\n",
      "Losses {'ner': 230.27171981334686}\n",
      "Losses {'ner': 225.57103788852692}\n",
      "Losses {'ner': 223.9265580177307}\n",
      "Losses {'ner': 217.44556975364685}\n",
      "Losses {'ner': 222.12992072105408}\n",
      "Losses {'ner': 217.72551822662354}\n",
      "Losses {'ner': 229.61383891105652}\n",
      "Losses {'ner': 211.38583654165268}\n",
      "Losses {'ner': 228.9752345085144}\n",
      "Losses {'ner': 217.15897274017334}\n",
      "Losses {'ner': 223.10197758674622}\n",
      "Losses {'ner': 231.56702363491058}\n",
      "Losses {'ner': 227.8536993265152}\n",
      "Losses {'ner': 229.99482142925262}\n",
      "Losses {'ner': 215.2111713886261}\n",
      "Losses {'ner': 209.4332573413849}\n",
      "Losses {'ner': 230.63561099767685}\n",
      "Losses {'ner': 229.44061887264252}\n",
      "Losses {'ner': 225.73978453874588}\n",
      "Losses {'ner': 213.96241199970245}\n",
      "Losses {'ner': 225.79383873939514}\n",
      "Losses {'ner': 225.07853865623474}\n",
      "Losses {'ner': 226.13606941699982}\n",
      "Losses {'ner': 214.6621254682541}\n",
      "Losses {'ner': 221.34867095947266}\n",
      "Losses {'ner': 209.84226644039154}\n",
      "Losses {'ner': 218.91071557998657}\n",
      "Losses {'ner': 208.3718421459198}\n",
      "Losses {'ner': 229.16357827186584}\n",
      "Losses {'ner': 219.03877782821655}\n",
      "Losses {'ner': 214.45256531238556}\n",
      "Losses {'ner': 226.73754382133484}\n",
      "Losses {'ner': 224.9558072090149}\n",
      "Losses {'ner': 227.36355590820312}\n",
      "Losses {'ner': 223.75510573387146}\n",
      "Losses {'ner': 212.71104745194316}\n",
      "Losses {'ner': 207.60406929254532}\n",
      "Losses {'ner': 220.52870672941208}\n",
      "Losses {'ner': 220.8569108247757}\n",
      "Losses {'ner': 227.8463351726532}\n",
      "Losses {'ner': 220.32000184059143}\n",
      "Losses {'ner': 237.21864247322083}\n",
      "Losses {'ner': 223.38902032375336}\n",
      "Losses {'ner': 212.51797258853912}\n",
      "Losses {'ner': 223.1768078804016}\n",
      "Losses {'ner': 218.48908495903015}\n",
      "Losses {'ner': 230.86435842514038}\n",
      "Losses {'ner': 207.96715784072876}\n",
      "Losses {'ner': 230.34611701965332}\n",
      "Losses {'ner': 226.57004380226135}\n",
      "Losses {'ner': 226.67022228240967}\n",
      "Losses {'ner': 215.2404704093933}\n",
      "Losses {'ner': 216.28339356184006}\n",
      "Losses {'ner': 216.3138643503189}\n",
      "Losses {'ner': 231.70960795879364}\n",
      "Losses {'ner': 214.53382515907288}\n",
      "Losses {'ner': 218.78469705581665}\n",
      "Losses {'ner': 220.2147946357727}\n",
      "Losses {'ner': 218.8448944091797}\n",
      "Losses {'ner': 215.98407113552094}\n",
      "Losses {'ner': 226.6694724559784}\n",
      "Losses {'ner': 226.0667645931244}\n",
      "Losses {'ner': 213.07185077667236}\n",
      "Losses {'ner': 224.9521360397339}\n",
      "Losses {'ner': 211.65704774856567}\n",
      "Losses {'ner': 222.68959951400757}\n",
      "Losses {'ner': 212.9945273399353}\n",
      "Losses {'ner': 225.48803567886353}\n",
      "Losses {'ner': 225.65701150894165}\n",
      "Losses {'ner': 231.84072637557983}\n",
      "Losses {'ner': 229.84266471862793}\n",
      "Losses {'ner': 215.46865713596344}\n",
      "Losses {'ner': 235.5315854549408}\n",
      "Losses {'ner': 218.2315719127655}\n",
      "Losses {'ner': 231.1956717967987}\n",
      "Losses {'ner': 213.41272377967834}\n",
      "Losses {'ner': 220.62911081314087}\n",
      "Losses {'ner': 218.87951707839966}\n",
      "Losses {'ner': 234.3837571144104}\n",
      "Losses {'ner': 214.13595926761627}\n",
      "Losses {'ner': 201.7188117802143}\n",
      "Losses {'ner': 217.07803416252136}\n",
      "Losses {'ner': 224.46319770812988}\n",
      "Losses {'ner': 222.34648370742798}\n",
      "Losses {'ner': 215.4781841635704}\n",
      "Losses {'ner': 232.49915981292725}\n",
      "Losses {'ner': 212.16674542427063}\n",
      "Losses {'ner': 216.21679151058197}\n",
      "Losses {'ner': 226.19832277297974}\n",
      "Losses {'ner': 233.07685923576355}\n",
      "Losses {'ner': 222.50607752799988}\n",
      "Losses {'ner': 216.3271622657776}\n",
      "Losses {'ner': 226.46128034591675}\n",
      "Losses {'ner': 229.93516767024994}\n",
      "Losses {'ner': 215.55992007255554}\n",
      "Losses {'ner': 222.5493643283844}\n",
      "Losses {'ner': 227.97429990768433}\n",
      "Losses {'ner': 210.51001852750778}\n",
      "Losses {'ner': 205.79398864507675}\n",
      "Losses {'ner': 224.5234158039093}\n",
      "Losses {'ner': 217.0300533771515}\n",
      "Losses {'ner': 211.04052591323853}\n",
      "Losses {'ner': 226.96336781978607}\n",
      "Losses {'ner': 226.53564608097076}\n",
      "Losses {'ner': 206.6495885848999}\n",
      "Losses {'ner': 195.7971675284207}\n",
      "Losses {'ner': 230.81073439121246}\n",
      "Losses {'ner': 223.0719131231308}\n",
      "Losses {'ner': 208.21593761444092}\n",
      "Losses {'ner': 219.72124993801117}\n",
      "Losses {'ner': 219.86478611826897}\n",
      "Losses {'ner': 221.31746566295624}\n",
      "Losses {'ner': 204.06475222110748}\n",
      "Losses {'ner': 230.56573963165283}\n",
      "Losses {'ner': 222.05735516548157}\n",
      "Losses {'ner': 235.00827026367188}\n",
      "Losses {'ner': 219.55795693397522}\n",
      "Losses {'ner': 224.0089703798294}\n",
      "Losses {'ner': 220.62684738636017}\n",
      "Losses {'ner': 209.06624138355255}\n",
      "Losses {'ner': 218.82647156715393}\n",
      "Losses {'ner': 221.49243187904358}\n",
      "Losses {'ner': 198.78052496910095}\n",
      "Losses {'ner': 233.430260181427}\n",
      "Losses {'ner': 223.72673189640045}\n",
      "Losses {'ner': 233.09541296958923}\n",
      "Losses {'ner': 222.6447458267212}\n",
      "Losses {'ner': 233.38916838169098}\n",
      "Losses {'ner': 226.99403476715088}\n",
      "Losses {'ner': 229.43764543533325}\n",
      "Losses {'ner': 222.3315805196762}\n",
      "Losses {'ner': 221.96220350265503}\n",
      "Losses {'ner': 219.9712073802948}\n",
      "Losses {'ner': 223.50513696670532}\n",
      "Losses {'ner': 221.29298269748688}\n",
      "Losses {'ner': 206.24098646640778}\n",
      "Losses {'ner': 219.21448242664337}\n",
      "Losses {'ner': 220.3049111366272}\n",
      "Losses {'ner': 219.6644183397293}\n",
      "Losses {'ner': 208.09659266471863}\n",
      "Losses {'ner': 226.2499579191208}\n",
      "Losses {'ner': 223.01984280347824}\n",
      "Losses {'ner': 212.81140732765198}\n",
      "Losses {'ner': 223.1851761341095}\n",
      "Losses {'ner': 226.41108226776123}\n",
      "Losses {'ner': 215.25215005874634}\n",
      "Losses {'ner': 216.22966039180756}\n",
      "Losses {'ner': 210.76745295524597}\n",
      "Losses {'ner': 222.98457086086273}\n",
      "Losses {'ner': 217.96201992034912}\n",
      "Losses {'ner': 225.85358810424805}\n",
      "Losses {'ner': 222.01921898126602}\n",
      "Losses {'ner': 230.4131133556366}\n",
      "Losses {'ner': 222.39131474494934}\n",
      "Losses {'ner': 219.14386796951294}\n",
      "Losses {'ner': 207.65529850125313}\n",
      "Losses {'ner': 219.3372106552124}\n",
      "Losses {'ner': 206.71452295780182}\n",
      "Losses {'ner': 218.88041424751282}\n",
      "Losses {'ner': 223.86933839321136}\n",
      "Losses {'ner': 215.0963693857193}\n",
      "Losses {'ner': 211.4606889486313}\n",
      "Losses {'ner': 218.7388846874237}\n",
      "Losses {'ner': 219.1254767179489}\n",
      "Losses {'ner': 200.7220757007599}\n",
      "Losses {'ner': 215.06743928790092}\n",
      "Losses {'ner': 211.74804812669754}\n",
      "Losses {'ner': 213.0709012746811}\n",
      "Losses {'ner': 216.36403894424438}\n",
      "Losses {'ner': 220.85530257225037}\n",
      "Losses {'ner': 208.86728048324585}\n",
      "Losses {'ner': 220.74122548103333}\n",
      "Losses {'ner': 217.57308292388916}\n",
      "Losses {'ner': 218.9680209159851}\n",
      "Losses {'ner': 212.27683281898499}\n",
      "Losses {'ner': 223.6740207672119}\n",
      "Losses {'ner': 238.3405830860138}\n",
      "Losses {'ner': 213.36580896377563}\n",
      "Losses {'ner': 207.6850802898407}\n",
      "Losses {'ner': 228.02270221710205}\n",
      "Losses {'ner': 217.8594274520874}\n",
      "Losses {'ner': 210.7708226442337}\n",
      "Losses {'ner': 211.78858709335327}\n",
      "Losses {'ner': 221.73332810401917}\n",
      "Losses {'ner': 214.55110454559326}\n",
      "Losses {'ner': 220.25524520874023}\n",
      "Losses {'ner': 217.92695033550262}\n",
      "Losses {'ner': 224.72950291633606}\n",
      "Losses {'ner': 215.4725695848465}\n",
      "Losses {'ner': 229.5975317955017}\n",
      "Losses {'ner': 209.9113422036171}\n",
      "Losses {'ner': 210.91896176338196}\n",
      "Losses {'ner': 229.90829753875732}\n",
      "Losses {'ner': 221.11987447738647}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Losses {'ner': 214.56935447454453}\n",
      "Losses {'ner': 231.15349847078323}\n",
      "Losses {'ner': 235.1015980243683}\n",
      "Losses {'ner': 223.11627650260925}\n",
      "Losses {'ner': 220.2607625722885}\n",
      "Losses {'ner': 224.77742564678192}\n",
      "Losses {'ner': 228.91698217391968}\n",
      "Losses {'ner': 236.9681648015976}\n",
      "Losses {'ner': 214.66164362430573}\n",
      "Losses {'ner': 221.76855039596558}\n",
      "Losses {'ner': 225.1822383403778}\n",
      "Losses {'ner': 213.94091254472733}\n",
      "Losses {'ner': 227.618816614151}\n",
      "Losses {'ner': 226.42103815078735}\n",
      "Losses {'ner': 217.4436103105545}\n",
      "Losses {'ner': 211.69355583190918}\n",
      "Losses {'ner': 207.6754274368286}\n",
      "Losses {'ner': 226.00589966773987}\n",
      "Losses {'ner': 212.1597522497177}\n",
      "Losses {'ner': 209.62838304042816}\n",
      "Losses {'ner': 219.4912281036377}\n",
      "Losses {'ner': 197.47383666038513}\n",
      "Losses {'ner': 221.9076075553894}\n",
      "Losses {'ner': 226.9108018875122}\n",
      "Losses {'ner': 212.63174188137054}\n",
      "Losses {'ner': 217.6819385290146}\n",
      "Losses {'ner': 221.81545159220695}\n",
      "Losses {'ner': 206.03863936662674}\n",
      "Losses {'ner': 200.90795934200287}\n",
      "Losses {'ner': 229.88366276025772}\n",
      "Losses {'ner': 218.23616480827332}\n",
      "Losses {'ner': 218.06837213039398}\n",
      "Losses {'ner': 214.0265326499939}\n",
      "Losses {'ner': 218.18335330486298}\n",
      "Losses {'ner': 225.37835669517517}\n",
      "Losses {'ner': 214.12336802482605}\n",
      "Losses {'ner': 211.44361293315887}\n",
      "Losses {'ner': 225.58165454864502}\n",
      "Losses {'ner': 219.69363379478455}\n",
      "Losses {'ner': 200.4398936033249}\n",
      "Losses {'ner': 214.36191483214498}\n",
      "Losses {'ner': 214.15797263383865}\n",
      "Losses {'ner': 229.4961130619049}\n",
      "Losses {'ner': 222.18852585554123}\n",
      "Losses {'ner': 217.41012001037598}\n",
      "Losses {'ner': 207.95875257253647}\n",
      "Losses {'ner': 220.68540343642235}\n",
      "Losses {'ner': 218.6739017367363}\n",
      "Losses {'ner': 231.77718424797058}\n",
      "Losses {'ner': 207.31668162345886}\n",
      "Losses {'ner': 233.92219352722168}\n",
      "Losses {'ner': 218.38767647743225}\n",
      "Losses {'ner': 220.84058916568756}\n",
      "Losses {'ner': 222.94186091423035}\n",
      "Losses {'ner': 232.53768849372864}\n",
      "Losses {'ner': 215.35155737400055}\n",
      "Losses {'ner': 222.46141600608826}\n",
      "Losses {'ner': 216.7162184715271}\n",
      "Losses {'ner': 216.13615763187408}\n",
      "Losses {'ner': 233.54604065418243}\n",
      "Losses {'ner': 227.67684316635132}\n",
      "Losses {'ner': 224.77241894602776}\n",
      "Losses {'ner': 238.65250205993652}\n",
      "Losses {'ner': 219.96824312210083}\n",
      "Losses {'ner': 235.00234079360962}\n",
      "Losses {'ner': 225.26641082763672}\n",
      "Losses {'ner': 228.95432043075562}\n",
      "Losses {'ner': 214.23487842082977}\n",
      "Losses {'ner': 201.50684773921967}\n",
      "Losses {'ner': 210.84110164642334}\n",
      "Losses {'ner': 219.7131164073944}\n",
      "Losses {'ner': 216.96092128753662}\n",
      "Losses {'ner': 219.0194799900055}\n",
      "Losses {'ner': 213.91375625133514}\n",
      "Losses {'ner': 218.74154424667358}\n",
      "Losses {'ner': 225.27457106113434}\n",
      "Losses {'ner': 219.17479062080383}\n",
      "Losses {'ner': 214.1135332584381}\n",
      "Losses {'ner': 218.89938151836395}\n",
      "Losses {'ner': 216.2852063179016}\n",
      "Losses {'ner': 214.87245070934296}\n",
      "Losses {'ner': 221.37911754846573}\n",
      "Losses {'ner': 200.85892236232758}\n",
      "Losses {'ner': 204.2814507484436}\n",
      "Losses {'ner': 215.29387307167053}\n",
      "Losses {'ner': 217.18575620651245}\n",
      "Losses {'ner': 226.75341749191284}\n",
      "Losses {'ner': 222.59326922893524}\n",
      "Losses {'ner': 213.56313407421112}\n",
      "Losses {'ner': 234.3611936569214}\n",
      "Losses {'ner': 223.61895656585693}\n",
      "Losses {'ner': 223.84624481201172}\n",
      "Losses {'ner': 223.4032485485077}\n",
      "Losses {'ner': 212.79656565189362}\n",
      "Losses {'ner': 219.05582404136658}\n",
      "Losses {'ner': 210.5790979862213}\n",
      "Losses {'ner': 224.78935742378235}\n",
      "Losses {'ner': 223.40033531188965}\n",
      "Losses {'ner': 220.63169825077057}\n",
      "Losses {'ner': 213.03167736530304}\n",
      "Losses {'ner': 232.2365369796753}\n",
      "Losses {'ner': 226.1428189277649}\n",
      "Losses {'ner': 210.10720133781433}\n",
      "Losses {'ner': 213.20139944553375}\n",
      "Losses {'ner': 209.95814406871796}\n",
      "Losses {'ner': 220.55750119686127}\n",
      "Losses {'ner': 233.4797866344452}\n",
      "Losses {'ner': 219.15956699848175}\n",
      "Losses {'ner': 222.03558206558228}\n",
      "Losses {'ner': 237.82080960273743}\n",
      "Losses {'ner': 211.5186386704445}\n",
      "Losses {'ner': 220.62300825119019}\n",
      "Losses {'ner': 205.33240443468094}\n",
      "Losses {'ner': 214.8964560031891}\n",
      "Losses {'ner': 211.92027723789215}\n",
      "Losses {'ner': 222.0203824043274}\n",
      "Losses {'ner': 209.84667670726776}\n",
      "Losses {'ner': 211.6052634716034}\n",
      "Losses {'ner': 220.90817546844482}\n",
      "Losses {'ner': 223.2134302854538}\n",
      "Losses {'ner': 213.76597905158997}\n",
      "Losses {'ner': 223.28703133761883}\n",
      "Losses {'ner': 203.2716028690338}\n",
      "Losses {'ner': 204.15361428260803}\n",
      "Losses {'ner': 228.4435208439827}\n",
      "Losses {'ner': 211.42567765712738}\n",
      "Losses {'ner': 215.6664433479309}\n",
      "Losses {'ner': 211.15016400814056}\n",
      "Losses {'ner': 205.95644092559814}\n",
      "Losses {'ner': 219.55003666877747}\n",
      "Losses {'ner': 222.5105916261673}\n",
      "Losses {'ner': 225.32554733753204}\n",
      "Losses {'ner': 210.09085965156555}\n",
      "Losses {'ner': 210.1142919063568}\n",
      "Losses {'ner': 221.25361323356628}\n",
      "Losses {'ner': 219.33860683441162}\n",
      "Losses {'ner': 212.4790107011795}\n",
      "Losses {'ner': 227.7150857448578}\n",
      "Losses {'ner': 217.35823607444763}\n",
      "Losses {'ner': 222.09060072898865}\n",
      "Losses {'ner': 215.31987178325653}\n",
      "Losses {'ner': 211.86396396160126}\n",
      "Losses {'ner': 230.3577492237091}\n",
      "Losses {'ner': 225.59962821006775}\n",
      "Losses {'ner': 207.92542445659637}\n",
      "Losses {'ner': 208.94251942634583}\n",
      "Losses {'ner': 218.34761214256287}\n",
      "Losses {'ner': 221.68787395954132}\n",
      "Losses {'ner': 210.07906806468964}\n",
      "Losses {'ner': 227.16365790367126}\n",
      "Losses {'ner': 215.68898582458496}\n",
      "Losses {'ner': 229.44952929019928}\n",
      "Losses {'ner': 229.65184020996094}\n",
      "Losses {'ner': 225.3011463880539}\n",
      "Losses {'ner': 209.66085493564606}\n",
      "Losses {'ner': 216.1791763305664}\n",
      "Losses {'ner': 222.88833022117615}\n",
      "Losses {'ner': 217.72879099845886}\n",
      "Losses {'ner': 218.4215636253357}\n",
      "Losses {'ner': 224.07511568069458}\n",
      "Losses {'ner': 221.29326391220093}\n",
      "Losses {'ner': 207.17542034387589}\n",
      "Losses {'ner': 216.36203587055206}\n",
      "Losses {'ner': 218.5512612760067}\n",
      "Losses {'ner': 216.61935412883759}\n",
      "Losses {'ner': 217.19040381908417}\n",
      "Losses {'ner': 219.75150084495544}\n",
      "Losses {'ner': 224.80935215950012}\n",
      "Losses {'ner': 216.52640008926392}\n",
      "Losses {'ner': 218.9731981754303}\n",
      "Losses {'ner': 226.58474361896515}\n",
      "Losses {'ner': 200.67816281318665}\n",
      "Losses {'ner': 225.3200261592865}\n",
      "Losses {'ner': 200.97563004493713}\n",
      "Losses {'ner': 227.68826580047607}\n",
      "Losses {'ner': 220.58463311195374}\n",
      "Losses {'ner': 212.27378916740417}\n",
      "Losses {'ner': 214.24386739730835}\n",
      "Losses {'ner': 215.9335799217224}\n",
      "Losses {'ner': 209.4794932603836}\n",
      "Losses {'ner': 211.79959857463837}\n",
      "Losses {'ner': 211.42180943489075}\n",
      "Losses {'ner': 211.90520071983337}\n",
      "Losses {'ner': 222.7405869960785}\n",
      "Losses {'ner': 224.36888766288757}\n",
      "Losses {'ner': 209.16932851076126}\n",
      "Losses {'ner': 225.51084065437317}\n",
      "Losses {'ner': 207.75216794013977}\n",
      "Losses {'ner': 234.78570091724396}\n",
      "Losses {'ner': 222.17448556423187}\n",
      "Losses {'ner': 214.28985488414764}\n",
      "Losses {'ner': 214.738147854805}\n",
      "Losses {'ner': 228.5299108028412}\n",
      "Losses {'ner': 227.90207076072693}\n",
      "Losses {'ner': 222.98458170890808}\n",
      "Losses {'ner': 219.60170531272888}\n",
      "Entities [('dabrafenib', 'gene'), ('trametinib', 'drug'), ('BRAF', 'gene'), ('V600', 'varient'), ('melanoma', 'diesease')]\n",
      "Tokens [('Thus', '', 2), (',', '', 2), ('dabrafenib', 'gene', 3), ('plus', '', 2), ('trametinib', 'drug', 3), ('provides', '', 2), ('an', '', 2), ('important', '', 2), ('treatment', '', 2), ('option', '', 2), ('for', '', 2), ('patients', '', 2), ('with', '', 2), ('BRAF', 'gene', 3), ('(', '', 2), ('V600', 'varient', 3), (')', '', 2), ('mutation', '', 2), ('-', '', 2), ('positive', '', 2), ('unresectable', '', 2), ('or', '', 2), ('metastatic', '', 2), ('melanoma', 'diesease', 3), ('.', '', 2)]\n",
      "Entities [('dabrafenib', 'gene'), ('vemurafenib', 'drug')]\n",
      "Tokens [('Combination', '', 2), ('therapy', '', 2), ('did', '', 2), ('not', '', 2), ('increase', '', 2), ('overall', '', 2), ('toxicity', '', 2), ('relative', '', 2), ('to', '', 2), ('dabrafenib', 'gene', 3), ('or', '', 2), ('vemurafenib', 'drug', 3), ('monotherapy', '', 2), (',', '', 2), ('with', '', 2), ('most', '', 2), ('adverse', '', 2), ('events', '', 2), ('(', '', 2), ('AEs', '', 2), (')', '', 2), ('mild', '', 2), ('or', '', 2), ('moderate', '', 2), ('in', '', 2), ('severity', '', 2), ('and', '', 2), ('generally', '', 2), ('manageable', '', 2), ('.', '', 2)]\n",
      "Entities [('dabrafenib and', 'LAW'), ('trametinib', 'drug')]\n",
      "Tokens [('This', '', 2), ('article', '', 2), ('reviews', '', 2), ('the', '', 2), ('therapeutic', '', 2), ('efficacy', '', 2), ('and', '', 2), ('tolerability', '', 2), ('of', '', 2), ('combination', '', 2), ('treatment', '', 2), ('with', '', 2), ('dabrafenib', 'LAW', 3), ('and', 'LAW', 1), ('trametinib', 'drug', 3), ('in', '', 2), ('this', '', 2), ('indication', '', 2), ('and', '', 2), ('summarizes', '', 2), ('relevant', '', 2), ('pharmacological', '', 2), ('data', '', 2), ('.', '', 2)]\n",
      "Entities [('dabrafenib or', 'LAW'), ('vemurafenib', 'drug')]\n",
      "Tokens [('Fewer', '', 2), ('skin', '', 2), ('-', '', 2), ('related', '', 2), ('AEs', '', 2), ('(', '', 2), ('e.g.', '', 2), ('cutaneous', '', 2), ('malignancies', '', 2), (',', '', 2), ('hyperkeratinosis', '', 2), ('and', '', 2), ('hand', '', 2), ('-', '', 2), ('foot', '', 2), ('syndrome', '', 2), (')', '', 2), ('were', '', 2), ('reported', '', 2), ('with', '', 2), ('combination', '', 2), ('therapy', '', 2), ('than', '', 2), ('with', '', 2), ('dabrafenib', 'LAW', 3), ('or', 'LAW', 1), ('vemurafenib', 'drug', 3), (',', '', 2), ('probably', '', 2), ('because', '', 2), ('of', '', 2), ('reduced', '', 2), ('paradoxical', '', 2), ('activation', '', 2), ('of', '', 2), ('the', '', 2), ('MAPK', '', 2), ('pathway', '', 2), ('.', '', 2)]\n",
      "Entities [('BRAF', 'gene'), ('dabrafenib', 'drug'), ('MEK', 'protien'), ('trametinib', 'drug'), ('melanoma', 'diesease'), ('BRAF', 'gene'), ('V600', 'varient')]\n",
      "Tokens [('The', '', 2), ('BRAF', 'gene', 3), ('inhibitor', '', 2), ('dabrafenib', 'drug', 3), ('(', '', 2), ('Tafinlar(Â', '', 2), ('®', '', 2), (')', '', 2), (')', '', 2), ('and', '', 2), ('the', '', 2), ('MEK', 'protien', 3), ('inhibitor', '', 2), ('trametinib', 'drug', 3), ('(', '', 2), ('Mekinist(Â', '', 2), ('®', '', 2), (')', '', 2), (')', '', 2), ('are', '', 2), ('indicated', '', 2), (',', '', 2), ('as', '', 2), ('monotherapy', '', 2), ('or', '', 2), ('in', '', 2), ('combination', '', 2), ('with', '', 2), ('each', '', 2), ('other', '', 2), (',', '', 2), ('for', '', 2), ('the', '', 2), ('treatment', '', 2), ('of', '', 2), ('patients', '', 2), ('with', '', 2), ('unresectable', '', 2), ('or', '', 2), ('metastatic', '', 2), ('melanoma', 'diesease', 3), ('with', '', 2), ('a', '', 2), ('BRAF', 'gene', 3), ('(', '', 2), ('V600', 'varient', 3), (')', '', 2), ('mutation', '', 2), ('.', '', 2)]\n",
      "Entities []\n",
      "Tokens [('Limited', '', 2), ('treatment', '', 2), ('benefit', '', 2), ('with', '', 2), ('the', '', 2), ('combination', '', 2), ('was', '', 2), ('also', '', 2), ('seen', '', 2), ('in', '', 2), ('patients', '', 2), ('who', '', 2), ('had', '', 2), ('progressed', '', 2), ('on', '', 2), ('prior', '', 2), ('BRAF', '', 2), ('inhibitor', '', 2), ('therapy', '', 2), (',', '', 2), ('as', '', 2), ('indicated', '', 2), ('by', '', 2), ('ORRs', '', 2), ('of', '', 2), ('â‰¤', '', 2), ('15Â', '', 2), ('%', '', 2), ('and', '', 2), ('stable', '', 2), ('disease', '', 2), ('in', '', 2), ('â‰¤', '', 2), ('50Â', '', 2), ('%', '', 2), ('of', '', 2), ('patients', '', 2), ('in', '', 2), ('small', '', 2), ('phase', '', 2), ('I', '', 2), ('and', '', 2), ('II', '', 2), ('studies', '', 2), ('.', '', 2), ('Limited', '', 2), ('treatment', '', 2), ('benefit', '', 2), ('with', '', 2), ('the', '', 2), ('combination', '', 2), ('was', '', 2), ('also', '', 2), ('seen', '', 2), ('in', '', 2), ('patients', '', 2), ('who', '', 2), ('had', '', 2), ('progressed', '', 2), ('on', '', 2), ('prior', '', 2), ('BRAF', '', 2), ('inhibitor', '', 2), ('therapy', '', 2), (',', '', 2), ('as', '', 2), ('indicated', '', 2), ('by', '', 2), ('ORRs', '', 2), ('of', '', 2), ('â‰¤', '', 2), ('15Â', '', 2), ('%', '', 2), ('and', '', 2), ('stable', '', 2), ('disease', '', 2), ('in', '', 2), ('â‰¤', '', 2), ('50Â', '', 2), ('%', '', 2), ('of', '', 2), ('patients', '', 2), ('in', '', 2), ('small', '', 2), ('phase', '', 2), ('I', '', 2), ('and', '', 2), ('II', '', 2), ('studies', '', 2), ('.', '', 2)]\n",
      "Entities [('Dabrafenib', 'drug'), ('trametinib', 'drug'), ('melanoma', 'diesease'), ('BRAF', 'gene'), ('V600E/K', 'varient')]\n",
      "Tokens [('Dabrafenib', 'drug', 3), ('plus', '', 2), ('trametinib', 'drug', 3), ('significantly', '', 2), ('prolonged', '', 2), ('progression', '', 2), ('-', '', 2), ('free', '', 2), ('survival', '', 2), ('(', '', 2), ('PFS', '', 2), (')', '', 2), ('and', '', 2), ('overall', '', 2), ('survival', '', 2), ('(', '', 2), ('OS', '', 2), (')', '', 2), (',', '', 2), ('improved', '', 2), ('objective', '', 2), ('response', '', 2), ('rates', '', 2), ('(', '', 2), ('ORRs', '', 2), (')', '', 2), ('and', '', 2), ('preserved', '', 2), ('health', '', 2), ('-', '', 2), ('related', '', 2), ('quality', '', 2), ('of', '', 2), ('life', '', 2), ('(', '', 2), ('HR', '', 2), ('-', '', 2), ('QOL', '', 2), (')', '', 2), ('to', '', 2), ('a', '', 2), ('greater', '', 2), ('extent', '', 2), ('than', '', 2), ('dabrafenib', '', 2), ('(', '', 2), ('in', '', 2), ('the', '', 2), ('double', '', 2), ('-', '', 2), ('blind', '', 2), ('COMBI', '', 2), ('-', '', 2), ('d', '', 2), ('study', '', 2), (')', '', 2), ('and', '', 2), ('vemurafenib', '', 2), ('(', '', 2), ('in', '', 2), ('the', '', 2), ('open', '', 2), ('-', '', 2), ('label', '', 2), ('COMBI', '', 2), ('-', '', 2), ('v', '', 2), ('study', '', 2), (')', '', 2), ('in', '', 2), ('two', '', 2), ('large', '', 2), (',', '', 2), ('randomized', '', 2), (',', '', 2), ('phase', '', 2), ('III', '', 2), ('studies', '', 2), ('in', '', 2), ('treatment', '', 2), ('-', '', 2), ('naÃ¯ve', '', 2), ('patients', '', 2), ('with', '', 2), ('unresectable', '', 2), ('or', '', 2), ('metastatic', '', 2), ('melanoma', 'diesease', 3), ('with', '', 2), ('BRAF', 'gene', 3), ('(', '', 2), ('V600E', 'varient', 3), ('/', 'varient', 1), ('K', 'varient', 1), (')', '', 2), ('mutation', '', 2), ('.', '', 2)]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saved model to /home/ubuntu/Grakn/Spacy Play/spacy_en_core_web_lg_V1.1\n",
      "Loading from /home/ubuntu/Grakn/Spacy Play/spacy_en_core_web_lg_V1.1\n",
      "Entities [('dabrafenib', 'gene'), ('trametinib', 'drug'), ('BRAF', 'gene'), ('V600', 'varient'), ('melanoma', 'diesease')]\n",
      "Tokens [('Thus', '', 2), (',', '', 2), ('dabrafenib', 'gene', 3), ('plus', '', 2), ('trametinib', 'drug', 3), ('provides', '', 2), ('an', '', 2), ('important', '', 2), ('treatment', '', 2), ('option', '', 2), ('for', '', 2), ('patients', '', 2), ('with', '', 2), ('BRAF', 'gene', 3), ('(', '', 2), ('V600', 'varient', 3), (')', '', 2), ('mutation', '', 2), ('-', '', 2), ('positive', '', 2), ('unresectable', '', 2), ('or', '', 2), ('metastatic', '', 2), ('melanoma', 'diesease', 3), ('.', '', 2)]\n",
      "Entities [('dabrafenib', 'gene'), ('vemurafenib', 'drug')]\n",
      "Tokens [('Combination', '', 2), ('therapy', '', 2), ('did', '', 2), ('not', '', 2), ('increase', '', 2), ('overall', '', 2), ('toxicity', '', 2), ('relative', '', 2), ('to', '', 2), ('dabrafenib', 'gene', 3), ('or', '', 2), ('vemurafenib', 'drug', 3), ('monotherapy', '', 2), (',', '', 2), ('with', '', 2), ('most', '', 2), ('adverse', '', 2), ('events', '', 2), ('(', '', 2), ('AEs', '', 2), (')', '', 2), ('mild', '', 2), ('or', '', 2), ('moderate', '', 2), ('in', '', 2), ('severity', '', 2), ('and', '', 2), ('generally', '', 2), ('manageable', '', 2), ('.', '', 2)]\n",
      "Entities [('dabrafenib and', 'LAW'), ('trametinib', 'drug')]\n",
      "Tokens [('This', '', 2), ('article', '', 2), ('reviews', '', 2), ('the', '', 2), ('therapeutic', '', 2), ('efficacy', '', 2), ('and', '', 2), ('tolerability', '', 2), ('of', '', 2), ('combination', '', 2), ('treatment', '', 2), ('with', '', 2), ('dabrafenib', 'LAW', 3), ('and', 'LAW', 1), ('trametinib', 'drug', 3), ('in', '', 2), ('this', '', 2), ('indication', '', 2), ('and', '', 2), ('summarizes', '', 2), ('relevant', '', 2), ('pharmacological', '', 2), ('data', '', 2), ('.', '', 2)]\n",
      "Entities [('dabrafenib or', 'LAW'), ('vemurafenib', 'drug')]\n",
      "Tokens [('Fewer', '', 2), ('skin', '', 2), ('-', '', 2), ('related', '', 2), ('AEs', '', 2), ('(', '', 2), ('e.g.', '', 2), ('cutaneous', '', 2), ('malignancies', '', 2), (',', '', 2), ('hyperkeratinosis', '', 2), ('and', '', 2), ('hand', '', 2), ('-', '', 2), ('foot', '', 2), ('syndrome', '', 2), (')', '', 2), ('were', '', 2), ('reported', '', 2), ('with', '', 2), ('combination', '', 2), ('therapy', '', 2), ('than', '', 2), ('with', '', 2), ('dabrafenib', 'LAW', 3), ('or', 'LAW', 1), ('vemurafenib', 'drug', 3), (',', '', 2), ('probably', '', 2), ('because', '', 2), ('of', '', 2), ('reduced', '', 2), ('paradoxical', '', 2), ('activation', '', 2), ('of', '', 2), ('the', '', 2), ('MAPK', '', 2), ('pathway', '', 2), ('.', '', 2)]\n",
      "Entities [('BRAF', 'gene'), ('dabrafenib', 'drug'), ('MEK', 'protien'), ('trametinib', 'drug'), ('melanoma', 'diesease'), ('BRAF', 'gene'), ('V600', 'varient')]\n",
      "Tokens [('The', '', 2), ('BRAF', 'gene', 3), ('inhibitor', '', 2), ('dabrafenib', 'drug', 3), ('(', '', 2), ('Tafinlar(Â', '', 2), ('®', '', 2), (')', '', 2), (')', '', 2), ('and', '', 2), ('the', '', 2), ('MEK', 'protien', 3), ('inhibitor', '', 2), ('trametinib', 'drug', 3), ('(', '', 2), ('Mekinist(Â', '', 2), ('®', '', 2), (')', '', 2), (')', '', 2), ('are', '', 2), ('indicated', '', 2), (',', '', 2), ('as', '', 2), ('monotherapy', '', 2), ('or', '', 2), ('in', '', 2), ('combination', '', 2), ('with', '', 2), ('each', '', 2), ('other', '', 2), (',', '', 2), ('for', '', 2), ('the', '', 2), ('treatment', '', 2), ('of', '', 2), ('patients', '', 2), ('with', '', 2), ('unresectable', '', 2), ('or', '', 2), ('metastatic', '', 2), ('melanoma', 'diesease', 3), ('with', '', 2), ('a', '', 2), ('BRAF', 'gene', 3), ('(', '', 2), ('V600', 'varient', 3), (')', '', 2), ('mutation', '', 2), ('.', '', 2)]\n",
      "Entities []\n",
      "Tokens [('Limited', '', 2), ('treatment', '', 2), ('benefit', '', 2), ('with', '', 2), ('the', '', 2), ('combination', '', 2), ('was', '', 2), ('also', '', 2), ('seen', '', 2), ('in', '', 2), ('patients', '', 2), ('who', '', 2), ('had', '', 2), ('progressed', '', 2), ('on', '', 2), ('prior', '', 2), ('BRAF', '', 2), ('inhibitor', '', 2), ('therapy', '', 2), (',', '', 2), ('as', '', 2), ('indicated', '', 2), ('by', '', 2), ('ORRs', '', 2), ('of', '', 2), ('â‰¤', '', 2), ('15Â', '', 2), ('%', '', 2), ('and', '', 2), ('stable', '', 2), ('disease', '', 2), ('in', '', 2), ('â‰¤', '', 2), ('50Â', '', 2), ('%', '', 2), ('of', '', 2), ('patients', '', 2), ('in', '', 2), ('small', '', 2), ('phase', '', 2), ('I', '', 2), ('and', '', 2), ('II', '', 2), ('studies', '', 2), ('.', '', 2), ('Limited', '', 2), ('treatment', '', 2), ('benefit', '', 2), ('with', '', 2), ('the', '', 2), ('combination', '', 2), ('was', '', 2), ('also', '', 2), ('seen', '', 2), ('in', '', 2), ('patients', '', 2), ('who', '', 2), ('had', '', 2), ('progressed', '', 2), ('on', '', 2), ('prior', '', 2), ('BRAF', '', 2), ('inhibitor', '', 2), ('therapy', '', 2), (',', '', 2), ('as', '', 2), ('indicated', '', 2), ('by', '', 2), ('ORRs', '', 2), ('of', '', 2), ('â‰¤', '', 2), ('15Â', '', 2), ('%', '', 2), ('and', '', 2), ('stable', '', 2), ('disease', '', 2), ('in', '', 2), ('â‰¤', '', 2), ('50Â', '', 2), ('%', '', 2), ('of', '', 2), ('patients', '', 2), ('in', '', 2), ('small', '', 2), ('phase', '', 2), ('I', '', 2), ('and', '', 2), ('II', '', 2), ('studies', '', 2), ('.', '', 2)]\n",
      "Entities [('Dabrafenib', 'drug'), ('trametinib', 'drug'), ('melanoma', 'diesease'), ('BRAF', 'gene'), ('V600E/K', 'varient')]\n",
      "Tokens [('Dabrafenib', 'drug', 3), ('plus', '', 2), ('trametinib', 'drug', 3), ('significantly', '', 2), ('prolonged', '', 2), ('progression', '', 2), ('-', '', 2), ('free', '', 2), ('survival', '', 2), ('(', '', 2), ('PFS', '', 2), (')', '', 2), ('and', '', 2), ('overall', '', 2), ('survival', '', 2), ('(', '', 2), ('OS', '', 2), (')', '', 2), (',', '', 2), ('improved', '', 2), ('objective', '', 2), ('response', '', 2), ('rates', '', 2), ('(', '', 2), ('ORRs', '', 2), (')', '', 2), ('and', '', 2), ('preserved', '', 2), ('health', '', 2), ('-', '', 2), ('related', '', 2), ('quality', '', 2), ('of', '', 2), ('life', '', 2), ('(', '', 2), ('HR', '', 2), ('-', '', 2), ('QOL', '', 2), (')', '', 2), ('to', '', 2), ('a', '', 2), ('greater', '', 2), ('extent', '', 2), ('than', '', 2), ('dabrafenib', '', 2), ('(', '', 2), ('in', '', 2), ('the', '', 2), ('double', '', 2), ('-', '', 2), ('blind', '', 2), ('COMBI', '', 2), ('-', '', 2), ('d', '', 2), ('study', '', 2), (')', '', 2), ('and', '', 2), ('vemurafenib', '', 2), ('(', '', 2), ('in', '', 2), ('the', '', 2), ('open', '', 2), ('-', '', 2), ('label', '', 2), ('COMBI', '', 2), ('-', '', 2), ('v', '', 2), ('study', '', 2), (')', '', 2), ('in', '', 2), ('two', '', 2), ('large', '', 2), (',', '', 2), ('randomized', '', 2), (',', '', 2), ('phase', '', 2), ('III', '', 2), ('studies', '', 2), ('in', '', 2), ('treatment', '', 2), ('-', '', 2), ('naÃ¯ve', '', 2), ('patients', '', 2), ('with', '', 2), ('unresectable', '', 2), ('or', '', 2), ('metastatic', '', 2), ('melanoma', 'diesease', 3), ('with', '', 2), ('BRAF', 'gene', 3), ('(', '', 2), ('V600E', 'varient', 3), ('/', 'varient', 1), ('K', 'varient', 1), (')', '', 2), ('mutation', '', 2), ('.', '', 2)]\n"
     ]
    }
   ],
   "source": [
    "if __name__ == \"__main__\":\n",
    "    main(model = \"en_core_web_lg\",output_dir= path_model_custom)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# NER "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_doc = 'Thus, dabrafenib plus trametinib provides an important treatment option for patients with BRAF (V600) mutation-positive unresectable or metastatic melanoma.'\n",
    "#'The BRAF inhibitor dabrafenib (Tafinlar(Â®)) and the MEK inhibitor trametinib (Mekinist(Â®)) are indicated, as monotherapy or in combination with each other, for the treatment of patients with unresectable or metastatic melanoma with a BRAF (V600) mutation. This article reviews the therapeutic efficacy and tolerability of combination treatment with dabrafenib and trametinib in this indication and summarizes relevant pharmacological data. Dabrafenib plus trametinib significantly prolonged progression-free survival (PFS) and overall survival (OS), improved objective response rates (ORRs) and preserved health-related quality of life (HR-QOL) to a greater extent than dabrafenib (in the double-blind COMBI-d study) and vemurafenib (in the open-label COMBI-v study) in two large, randomized, phase III studies in treatment-naÃ¯ve patients with unresectable or metastatic melanoma with BRAF (V600E/K) mutation. Limited treatment benefit with the combination was also seen in patients who had progressed on prior BRAF inhibitor therapy, as indicated by ORRs of â‰¤ 15Â\\xa0% and stable disease in â‰¤ 50Â\\xa0% of patients in small phase I and II studies. Combination therapy did not increase overall toxicity relative to dabrafenib or vemurafenib monotherapy, with most adverse events (AEs) mild or moderate in severity and generally manageable. Fewer skin-related AEs (e.g. cutaneous malignancies, hyperkeratinosis and hand-foot syndrome) were reported with combination therapy than with dabrafenib or vemurafenib, probably because of reduced paradoxical activation of the MAPK pathway. Thus, dabrafenib plus trametinib provides an important treatment option for patients with BRAF (V600) mutation-positive unresectable or metastatic melanoma.'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dabrafenib --> gene\n",
      "trametinib --> drug\n",
      "BRAF --> gene\n",
      "V600 --> varient\n",
      "melanoma --> diesease\n"
     ]
    }
   ],
   "source": [
    "#nlp = spacy.load(\"en_core_web_lg\")\n",
    "#doc = nlp(u\"Apple is looking at buying U.K. startup for $1 billion\")\n",
    "\n",
    "nlp = spacy.load(path_model_custom)#(\"/home/ubuntu/Grakn/Spacy Play/spacy_en_core_web_lg\")\n",
    "doc1 = nlp(test_doc)\n",
    "\n",
    "nlp2 = spacy.load(\"/home/ubuntu/Grakn/Spacy Play/spacy_en_core_web_lg\")\n",
    "doc2 = nlp2(test_doc)\n",
    "\n",
    "\n",
    "# for ent in doc.ents:\n",
    "#     print(ent.text, ent.start_char, ent.end_char, ent.label_)\n",
    "\n",
    "for ent in doc1.ents:\n",
    "    print(ent.text,\"-->\",ent.label_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "trametinib --> drug\n",
      "BRAF --> gene\n",
      "V600 --> varient\n",
      "melanoma --> diesease\n"
     ]
    }
   ],
   "source": [
    "for ent in doc2.ents:\n",
    "    print(ent.text,\"-->\",ent.label_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Playing with Spacy Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "doc ='''The BRAF inhibitor dabrafenib (Tafinlar(Â®)) and the MEK inhibitor trametinib (Mekinist(Â®)) are indicated, as monotherapy or in combination with each other, for the treatment of patients with unresectable or metastatic melanoma with a BRAF (V600) mutation. \n",
    "This article reviews the therapeutic efficacy and tolerability of combination treatment with dabrafenib and trametinib in this indication and summarizes relevant pharmacological data. \n",
    "Dabrafenib plus trametinib significantly prolonged progression-free survival (PFS) and overall survival (OS), improved objective response rates (ORRs) and preserved health-related quality of life (HR-QOL) to a greater extent than dabrafenib (in the double-blind COMBI-d study) and vemurafenib (in the open-label COMBI-v study) in two large, randomized, phase III studies in treatment-naÃ¯ve patients with unresectable or metastatic melanoma with BRAF (V600E/K) mutation. Limited treatment benefit with the combination was also seen in patients who had progressed on prior BRAF inhibitor therapy, as indicated by ORRs of â‰¤ 15Â % and stable disease in â‰¤ 50Â % of patients in small phase I and II studies. Combination therapy did not increase overall toxicity relative to dabrafenib or vemurafenib monotherapy, with most adverse events (AEs) mild or moderate in severity and generally manageable. Fewer skin-related AEs (e.g. cutaneous malignancies, hyperkeratinosis and hand-foot syndrome) were reported with combination therapy than with dabrafenib or vemurafenib, probably because of reduced paradoxical activation of the MAPK pathway. Thus, dabrafenib plus trametinib provides an important treatment option for patients with BRAF (V600) mutation-positive unresectable or metastatic melanoma.\n",
    "'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "doc1 = '''Hi, my name is Tobias. I’m a trained mathematician that now works as a data scientist and machine learning engineer. In 2018 I achieved the status of a kaggle master and still spend quite some time there.'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Loading spacy model\n",
    "# In this case, trained model is being loaded\n",
    "nlp = spacy.load(\"/home/ubuntu/Grakn/Spacy Play/spacy_en_core_web_lg\")#(\"en_core_web_sm\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Parsing document to spacy loaded model\n",
    "nlp_doc = nlp(doc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "#---- POS Tagging ----#\n",
      "\n",
      "The --> the --> DET --> DT --> det --> Xxx --> True --> True\n",
      "BRAF --> BRAF --> PROPN --> NNP --> compound --> XXXX --> True --> False\n",
      "inhibitor --> inhibitor --> NOUN --> NN --> compound --> xxxx --> True --> False\n",
      "dabrafenib --> dabrafenib --> NOUN --> NN --> nsubjpass --> xxxx --> True --> False\n",
      "( --> ( --> PUNCT --> -LRB- --> punct --> ( --> False --> False\n",
      "Tafinlar(Â --> Tafinlar(Â --> PROPN --> NNP --> compound --> Xxxxx(X --> False --> False\n",
      "® --> ® --> PROPN --> NNP --> appos --> ® --> False --> False\n",
      ") --> ) --> PUNCT --> -RRB- --> punct --> ) --> False --> False\n",
      ") --> ) --> PUNCT --> -RRB- --> punct --> ) --> False --> False\n",
      "and --> and --> CCONJ --> CC --> cc --> xxx --> True --> True\n",
      "the --> the --> DET --> DT --> det --> xxx --> True --> True\n",
      "MEK --> MEK --> PROPN --> NNP --> compound --> XXX --> True --> False\n",
      "inhibitor --> inhibitor --> NOUN --> NN --> compound --> xxxx --> True --> False\n",
      "trametinib --> trametinib --> NOUN --> NN --> conj --> xxxx --> True --> False\n",
      "( --> ( --> PUNCT --> -LRB- --> punct --> ( --> False --> False\n",
      "Mekinist(Â --> Mekinist(Â --> PROPN --> NNP --> compound --> Xxxxx(X --> False --> False\n",
      "® --> ® --> PROPN --> NNP --> appos --> ® --> False --> False\n",
      ") --> ) --> PUNCT --> -RRB- --> punct --> ) --> False --> False\n",
      ") --> ) --> PUNCT --> -RRB- --> punct --> ) --> False --> False\n",
      "are --> be --> VERB --> VBP --> auxpass --> xxx --> True --> True\n",
      "indicated --> indicate --> VERB --> VBN --> ROOT --> xxxx --> True --> False\n",
      ", --> , --> PUNCT --> , --> punct --> , --> False --> False\n",
      "as --> as --> ADP --> IN --> advmod --> xx --> True --> True\n",
      "monotherapy --> monotherapy --> ADV --> RB --> advmod --> xxxx --> True --> False\n",
      "or --> or --> CCONJ --> CC --> cc --> xx --> True --> True\n",
      "in --> in --> ADP --> IN --> advcl --> xx --> True --> True\n",
      "combination --> combination --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      "with --> with --> ADP --> IN --> prep --> xxxx --> True --> True\n",
      "each --> each --> DET --> DT --> det --> xxxx --> True --> True\n",
      "other --> other --> ADJ --> JJ --> pobj --> xxxx --> True --> True\n",
      ", --> , --> PUNCT --> , --> punct --> , --> False --> False\n",
      "for --> for --> ADP --> IN --> conj --> xxx --> True --> True\n",
      "the --> the --> DET --> DT --> det --> xxx --> True --> True\n",
      "treatment --> treatment --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      "of --> of --> ADP --> IN --> prep --> xx --> True --> True\n",
      "patients --> patient --> NOUN --> NNS --> pobj --> xxxx --> True --> False\n",
      "with --> with --> ADP --> IN --> prep --> xxxx --> True --> True\n",
      "unresectable --> unresectable --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "or --> or --> CCONJ --> CC --> cc --> xx --> True --> True\n",
      "metastatic --> metastatic --> ADJ --> JJ --> conj --> xxxx --> True --> False\n",
      "melanoma --> melanoma --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      "with --> with --> ADP --> IN --> prep --> xxxx --> True --> True\n",
      "a --> a --> DET --> DT --> det --> x --> True --> True\n",
      "BRAF --> BRAF --> PROPN --> NNP --> nmod --> XXXX --> True --> False\n",
      "( --> ( --> PUNCT --> -LRB- --> punct --> ( --> False --> False\n",
      "V600 --> V600 --> PROPN --> NNP --> subtok --> Xddd --> False --> False\n",
      ") --> ) --> PUNCT --> -RRB- --> punct --> ) --> False --> False\n",
      "mutation --> mutation --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      ". --> . --> PUNCT --> . --> punct --> . --> False --> False\n",
      "\n",
      " --> \n",
      " --> SPACE --> _SP -->  --> \n",
      " --> False --> False\n",
      "This --> this --> DET --> DT --> det --> Xxxx --> True --> True\n",
      "article --> article --> NOUN --> NN --> nsubj --> xxxx --> True --> False\n",
      "reviews --> review --> VERB --> VBZ --> ROOT --> xxxx --> True --> False\n",
      "the --> the --> DET --> DT --> det --> xxx --> True --> True\n",
      "therapeutic --> therapeutic --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "efficacy --> efficacy --> NOUN --> NN --> dobj --> xxxx --> True --> False\n",
      "and --> and --> CCONJ --> CC --> cc --> xxx --> True --> True\n",
      "tolerability --> tolerability --> NOUN --> NN --> conj --> xxxx --> True --> False\n",
      "of --> of --> ADP --> IN --> prep --> xx --> True --> True\n",
      "combination --> combination --> NOUN --> NN --> compound --> xxxx --> True --> False\n",
      "treatment --> treatment --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      "with --> with --> ADP --> IN --> prep --> xxxx --> True --> True\n",
      "dabrafenib --> dabrafenib --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      "and --> and --> CCONJ --> CC --> cc --> xxx --> True --> True\n",
      "trametinib --> trametinib --> NOUN --> NN --> conj --> xxxx --> True --> False\n",
      "in --> in --> ADP --> IN --> prep --> xx --> True --> True\n",
      "this --> this --> DET --> DT --> det --> xxxx --> True --> True\n",
      "indication --> indication --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      "and --> and --> CCONJ --> CC --> cc --> xxx --> True --> True\n",
      "summarizes --> summarize --> VERB --> VBZ --> conj --> xxxx --> True --> False\n",
      "relevant --> relevant --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "pharmacological --> pharmacological --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "data --> datum --> NOUN --> NNS --> dobj --> xxxx --> True --> False\n",
      ". --> . --> PUNCT --> . --> punct --> . --> False --> False\n",
      "\n",
      " --> \n",
      " --> SPACE --> _SP -->  --> \n",
      " --> False --> False\n",
      "Dabrafenib --> Dabrafenib --> PROPN --> NNP --> nsubj --> Xxxxx --> True --> False\n",
      "plus --> plus --> CCONJ --> CC --> cc --> xxxx --> True --> False\n",
      "trametinib --> trametinib --> NOUN --> NN --> conj --> xxxx --> True --> False\n",
      "significantly --> significantly --> ADV --> RB --> advmod --> xxxx --> True --> False\n",
      "prolonged --> prolong --> VERB --> VBD --> amod --> xxxx --> True --> False\n",
      "progression --> progression --> NOUN --> NN --> npadvmod --> xxxx --> True --> False\n",
      "- --> - --> PUNCT --> HYPH --> punct --> - --> False --> False\n",
      "free --> free --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "survival --> survival --> NOUN --> NN --> conj --> xxxx --> True --> False\n",
      "( --> ( --> PUNCT --> -LRB- --> punct --> ( --> False --> False\n",
      "PFS --> PFS --> PROPN --> NNP --> appos --> XXX --> True --> False\n",
      ") --> ) --> PUNCT --> -RRB- --> punct --> ) --> False --> False\n",
      "and --> and --> CCONJ --> CC --> cc --> xxx --> True --> True\n",
      "overall --> overall --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "survival --> survival --> NOUN --> NN --> conj --> xxxx --> True --> False\n",
      "( --> ( --> PUNCT --> -LRB- --> punct --> ( --> False --> False\n",
      "OS --> OS --> PROPN --> NNP --> appos --> XX --> True --> False\n",
      ") --> ) --> PUNCT --> -RRB- --> punct --> ) --> False --> False\n",
      ", --> , --> PUNCT --> , --> punct --> , --> False --> False\n",
      "improved --> improve --> VERB --> VBN --> ROOT --> xxxx --> True --> False\n",
      "objective --> objective --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "response --> response --> NOUN --> NN --> compound --> xxxx --> True --> False\n",
      "rates --> rate --> NOUN --> NNS --> dobj --> xxxx --> True --> False\n",
      "( --> ( --> PUNCT --> -LRB- --> punct --> ( --> False --> False\n",
      "ORRs --> orr --> NOUN --> NNS --> appos --> XXXx --> True --> False\n",
      ") --> ) --> PUNCT --> -RRB- --> punct --> ) --> False --> False\n",
      "and --> and --> CCONJ --> CC --> cc --> xxx --> True --> True\n",
      "preserved --> preserve --> VERB --> VBN --> amod --> xxxx --> True --> False\n",
      "health --> health --> NOUN --> NN --> npadvmod --> xxxx --> True --> False\n",
      "- --> - --> PUNCT --> HYPH --> punct --> - --> False --> False\n",
      "related --> relate --> VERB --> VBN --> amod --> xxxx --> True --> False\n",
      "quality --> quality --> NOUN --> NN --> conj --> xxxx --> True --> False\n",
      "of --> of --> ADP --> IN --> prep --> xx --> True --> True\n",
      "life --> life --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      "( --> ( --> PUNCT --> -LRB- --> punct --> ( --> False --> False\n",
      "HR --> HR --> PROPN --> NNP --> compound --> XX --> True --> False\n",
      "- --> - --> PUNCT --> HYPH --> punct --> - --> False --> False\n",
      "QOL --> QOL --> PROPN --> NNP --> appos --> XXX --> True --> False\n",
      ") --> ) --> PUNCT --> -RRB- --> punct --> ) --> False --> False\n",
      "to --> to --> ADP --> IN --> prep --> xx --> True --> True\n",
      "a --> a --> DET --> DT --> det --> x --> True --> True\n",
      "greater --> great --> ADJ --> JJR --> amod --> xxxx --> True --> False\n",
      "extent --> extent --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      "than --> than --> ADP --> IN --> prep --> xxxx --> True --> True\n",
      "dabrafenib --> dabrafenib --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      "( --> ( --> PUNCT --> -LRB- --> punct --> ( --> False --> False\n",
      "in --> in --> ADP --> IN --> prep --> xx --> True --> True\n",
      "the --> the --> DET --> DT --> det --> xxx --> True --> True\n",
      "double --> double --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "- --> - --> PUNCT --> HYPH --> punct --> - --> False --> False\n",
      "blind --> blind --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "COMBI --> combi --> NOUN --> NN --> compound --> XXXX --> True --> False\n",
      "- --> - --> PUNCT --> HYPH --> punct --> - --> False --> False\n",
      "d --> d --> ADJ --> JJ --> compound --> x --> True --> False\n",
      "study --> study --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      ") --> ) --> PUNCT --> -RRB- --> punct --> ) --> False --> False\n",
      "and --> and --> CCONJ --> CC --> cc --> xxx --> True --> True\n",
      "vemurafenib --> vemurafenib --> NOUN --> NN --> conj --> xxxx --> True --> False\n",
      "( --> ( --> PUNCT --> -LRB- --> punct --> ( --> False --> False\n",
      "in --> in --> ADP --> IN --> prep --> xx --> True --> True\n",
      "the --> the --> DET --> DT --> det --> xxx --> True --> True\n",
      "open --> open --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "- --> - --> PUNCT --> HYPH --> punct --> - --> False --> False\n",
      "label --> label --> NOUN --> NN --> compound --> xxxx --> True --> False\n",
      "COMBI --> COMBI --> PROPN --> NNP --> compound --> XXXX --> True --> False\n",
      "- --> - --> PUNCT --> HYPH --> punct --> - --> False --> False\n",
      "v --> v --> PROPN --> NNP --> compound --> x --> True --> False\n",
      "study --> study --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      ") --> ) --> PUNCT --> -RRB- --> punct --> ) --> False --> False\n",
      "in --> in --> ADP --> IN --> prep --> xx --> True --> True\n",
      "two --> two --> NUM --> CD --> nummod --> xxx --> True --> True\n",
      "large --> large --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      ", --> , --> PUNCT --> , --> punct --> , --> False --> False\n",
      "randomized --> randomize --> VERB --> VBN --> conj --> xxxx --> True --> False\n",
      ", --> , --> PUNCT --> , --> punct --> , --> False --> False\n",
      "phase --> phase --> NOUN --> NN --> compound --> xxxx --> True --> False\n",
      "III --> iii --> NUM --> CD --> compound --> XXX --> True --> False\n",
      "studies --> study --> NOUN --> NNS --> pobj --> xxxx --> True --> False\n",
      "in --> in --> ADP --> IN --> prep --> xx --> True --> True\n",
      "treatment --> treatment --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      "- --> - --> PUNCT --> HYPH --> punct --> - --> False --> False\n",
      "naÃ¯ve --> naã¯ve --> NUM --> CD --> punct --> xxX¯xx --> False --> False\n",
      "patients --> patient --> NOUN --> NNS --> dobj --> xxxx --> True --> False\n",
      "with --> with --> ADP --> IN --> prep --> xxxx --> True --> True\n",
      "unresectable --> unresectable --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "or --> or --> CCONJ --> CC --> cc --> xx --> True --> True\n",
      "metastatic --> metastatic --> ADJ --> JJ --> conj --> xxxx --> True --> False\n",
      "melanoma --> melanoma --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      "with --> with --> ADP --> IN --> prep --> xxxx --> True --> True\n",
      "BRAF --> BRAF --> PROPN --> NNP --> nmod --> XXXX --> True --> False\n",
      "( --> ( --> PUNCT --> -LRB- --> punct --> ( --> False --> False\n",
      "V600E --> V600E --> PROPN --> NNP --> nmod --> XdddX --> False --> False\n",
      "/ --> / --> SYM --> SYM --> punct --> / --> False --> False\n",
      "K --> K --> PROPN --> NNP --> nmod --> X --> True --> False\n",
      ") --> ) --> PUNCT --> -RRB- --> punct --> ) --> False --> False\n",
      "mutation --> mutation --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      ". --> . --> PUNCT --> . --> punct --> . --> False --> False\n",
      "Limited --> limited --> ADJ --> JJ --> amod --> Xxxxx --> True --> False\n",
      "treatment --> treatment --> NOUN --> NN --> compound --> xxxx --> True --> False\n",
      "benefit --> benefit --> NOUN --> NN --> nsubjpass --> xxxx --> True --> False\n",
      "with --> with --> ADP --> IN --> prep --> xxxx --> True --> True\n",
      "the --> the --> DET --> DT --> det --> xxx --> True --> True\n",
      "combination --> combination --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      "was --> be --> VERB --> VBD --> auxpass --> xxx --> True --> True\n",
      "also --> also --> ADV --> RB --> advmod --> xxxx --> True --> True\n",
      "seen --> see --> VERB --> VBN --> ROOT --> xxxx --> True --> False\n",
      "in --> in --> ADP --> IN --> prep --> xx --> True --> True\n",
      "patients --> patient --> NOUN --> NNS --> pobj --> xxxx --> True --> False\n",
      "who --> who --> PRON --> WP --> nsubj --> xxx --> True --> True\n",
      "had --> have --> VERB --> VBD --> aux --> xxx --> True --> True\n",
      "progressed --> progress --> VERB --> VBN --> relcl --> xxxx --> True --> False\n",
      "on --> on --> ADP --> IN --> prep --> xx --> True --> True\n",
      "prior --> prior --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "BRAF --> BRAF --> PROPN --> NNP --> compound --> XXXX --> True --> False\n",
      "inhibitor --> inhibitor --> NOUN --> NN --> compound --> xxxx --> True --> False\n",
      "therapy --> therapy --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      ", --> , --> PUNCT --> , --> punct --> , --> False --> False\n",
      "as --> as --> ADP --> IN --> mark --> xx --> True --> True\n",
      "indicated --> indicate --> VERB --> VBN --> advcl --> xxxx --> True --> False\n",
      "by --> by --> ADP --> IN --> agent --> xx --> True --> True\n",
      "ORRs --> orr --> NOUN --> NNS --> pobj --> XXXx --> True --> False\n",
      "of --> of --> ADP --> IN --> prep --> xx --> True --> True\n",
      "â‰¤ --> â‰¤ --> SYM --> $ --> nmod --> x‰¤ --> False --> False\n",
      "15Â --> 15â --> NUM --> CD --> nummod --> ddX --> False --> False\n",
      "% --> % --> NOUN --> NN --> pobj --> % --> False --> False\n",
      "and --> and --> CCONJ --> CC --> cc --> xxx --> True --> True\n",
      "stable --> stable --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "disease --> disease --> NOUN --> NN --> conj --> xxxx --> True --> False\n",
      "in --> in --> ADP --> IN --> prep --> xx --> True --> True\n",
      "â‰¤ --> â‰¤ --> SYM --> $ --> pobj --> x‰¤ --> False --> False\n",
      "50Â --> 50â --> NUM --> CD --> nummod --> ddX --> False --> False\n",
      "% --> % --> NOUN --> NN --> dobj --> % --> False --> False\n",
      "of --> of --> ADP --> IN --> prep --> xx --> True --> True\n",
      "patients --> patient --> NOUN --> NNS --> pobj --> xxxx --> True --> False\n",
      "in --> in --> ADP --> IN --> prep --> xx --> True --> True\n",
      "small --> small --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "phase --> phase --> NOUN --> NN --> nmod --> xxxx --> True --> False\n",
      "I --> -PRON- --> PRON --> PRP --> nmod --> X --> True --> True\n",
      "and --> and --> CCONJ --> CC --> cc --> xxx --> True --> True\n",
      "II --> II --> PROPN --> NNP --> conj --> XX --> True --> False\n",
      "studies --> study --> NOUN --> NNS --> pobj --> xxxx --> True --> False\n",
      ". --> . --> PUNCT --> . --> punct --> . --> False --> False\n",
      "Combination --> combination --> NOUN --> NN --> compound --> Xxxxx --> True --> False\n",
      "therapy --> therapy --> NOUN --> NN --> nsubj --> xxxx --> True --> False\n",
      "did --> do --> VERB --> VBD --> aux --> xxx --> True --> True\n",
      "not --> not --> ADV --> RB --> neg --> xxx --> True --> True\n",
      "increase --> increase --> VERB --> VB --> ROOT --> xxxx --> True --> False\n",
      "overall --> overall --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "toxicity --> toxicity --> NOUN --> NN --> dobj --> xxxx --> True --> False\n",
      "relative --> relative --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "to --> to --> ADP --> IN --> prep --> xx --> True --> True\n",
      "dabrafenib --> dabrafenib --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      "or --> or --> CCONJ --> CC --> cc --> xx --> True --> True\n",
      "vemurafenib --> vemurafenib --> NOUN --> NN --> conj --> xxxx --> True --> False\n",
      "monotherapy --> monotherapy --> NOUN --> NN --> advmod --> xxxx --> True --> False\n",
      ", --> , --> PUNCT --> , --> punct --> , --> False --> False\n",
      "with --> with --> ADP --> IN --> prep --> xxxx --> True --> True\n",
      "most --> most --> ADJ --> JJS --> amod --> xxxx --> True --> True\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "adverse --> adverse --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "events --> event --> NOUN --> NNS --> pobj --> xxxx --> True --> False\n",
      "( --> ( --> PUNCT --> -LRB- --> punct --> ( --> False --> False\n",
      "AEs --> AEs --> PROPN --> NNP --> appos --> XXx --> True --> False\n",
      ") --> ) --> PUNCT --> -RRB- --> punct --> ) --> False --> False\n",
      "mild --> mild --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "or --> or --> CCONJ --> CC --> cc --> xx --> True --> True\n",
      "moderate --> moderate --> ADJ --> JJ --> conj --> xxxx --> True --> False\n",
      "in --> in --> ADP --> IN --> prep --> xx --> True --> True\n",
      "severity --> severity --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      "and --> and --> CCONJ --> CC --> cc --> xxx --> True --> True\n",
      "generally --> generally --> ADV --> RB --> advmod --> xxxx --> True --> False\n",
      "manageable --> manageable --> ADJ --> JJ --> conj --> xxxx --> True --> False\n",
      ". --> . --> PUNCT --> . --> punct --> . --> False --> False\n",
      "Fewer --> few --> ADJ --> JJR --> amod --> Xxxxx --> True --> False\n",
      "skin --> skin --> NOUN --> NN --> npadvmod --> xxxx --> True --> False\n",
      "- --> - --> PUNCT --> HYPH --> punct --> - --> False --> False\n",
      "related --> relate --> VERB --> VBN --> amod --> xxxx --> True --> False\n",
      "AEs --> ae --> NOUN --> NNS --> nsubjpass --> XXx --> True --> False\n",
      "( --> ( --> PUNCT --> -LRB- --> punct --> ( --> False --> False\n",
      "e.g. --> e.g. --> X --> FW --> amod --> x.x. --> False --> False\n",
      "cutaneous --> cutaneous --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "malignancies --> malignancy --> NOUN --> NNS --> appos --> xxxx --> True --> False\n",
      ", --> , --> PUNCT --> , --> punct --> , --> False --> False\n",
      "hyperkeratinosis --> hyperkeratinosis --> NOUN --> NN --> conj --> xxxx --> True --> False\n",
      "and --> and --> CCONJ --> CC --> cc --> xxx --> True --> True\n",
      "hand --> hand --> NOUN --> NN --> compound --> xxxx --> True --> False\n",
      "- --> - --> PUNCT --> HYPH --> punct --> - --> False --> False\n",
      "foot --> foot --> NOUN --> NN --> compound --> xxxx --> True --> False\n",
      "syndrome --> syndrome --> NOUN --> NN --> conj --> xxxx --> True --> False\n",
      ") --> ) --> PUNCT --> -RRB- --> punct --> ) --> False --> False\n",
      "were --> be --> VERB --> VBD --> auxpass --> xxxx --> True --> True\n",
      "reported --> report --> VERB --> VBN --> ROOT --> xxxx --> True --> False\n",
      "with --> with --> ADP --> IN --> prep --> xxxx --> True --> True\n",
      "combination --> combination --> NOUN --> NN --> compound --> xxxx --> True --> False\n",
      "therapy --> therapy --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      "than --> than --> ADP --> IN --> prep --> xxxx --> True --> True\n",
      "with --> with --> ADP --> IN --> prep --> xxxx --> True --> True\n",
      "dabrafenib --> dabrafenib --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      "or --> or --> CCONJ --> CC --> cc --> xx --> True --> True\n",
      "vemurafenib --> vemurafenib --> NOUN --> NN --> conj --> xxxx --> True --> False\n",
      ", --> , --> PUNCT --> , --> punct --> , --> False --> False\n",
      "probably --> probably --> ADV --> RB --> advmod --> xxxx --> True --> False\n",
      "because --> because --> ADP --> IN --> prep --> xxxx --> True --> True\n",
      "of --> of --> ADP --> IN --> pcomp --> xx --> True --> True\n",
      "reduced --> reduced --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "paradoxical --> paradoxical --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "activation --> activation --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      "of --> of --> ADP --> IN --> prep --> xx --> True --> True\n",
      "the --> the --> DET --> DT --> det --> xxx --> True --> True\n",
      "MAPK --> MAPK --> PROPN --> NNP --> compound --> XXXX --> True --> False\n",
      "pathway --> pathway --> NOUN --> NN --> pobj --> xxxx --> True --> False\n",
      ". --> . --> PUNCT --> . --> punct --> . --> False --> False\n",
      "Thus --> thus --> ADV --> RB --> advmod --> Xxxx --> True --> True\n",
      ", --> , --> PUNCT --> , --> punct --> , --> False --> False\n",
      "dabrafenib --> dabrafenib --> NOUN --> NN --> nsubj --> xxxx --> True --> False\n",
      "plus --> plus --> CCONJ --> CC --> cc --> xxxx --> True --> False\n",
      "trametinib --> trametinib --> NOUN --> NN --> conj --> xxxx --> True --> False\n",
      "provides --> provide --> VERB --> VBZ --> ROOT --> xxxx --> True --> False\n",
      "an --> an --> DET --> DT --> det --> xx --> True --> True\n",
      "important --> important --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "treatment --> treatment --> NOUN --> NN --> compound --> xxxx --> True --> False\n",
      "option --> option --> NOUN --> NN --> dobj --> xxxx --> True --> False\n",
      "for --> for --> ADP --> IN --> prep --> xxx --> True --> True\n",
      "patients --> patient --> NOUN --> NNS --> pobj --> xxxx --> True --> False\n",
      "with --> with --> ADP --> IN --> prep --> xxxx --> True --> True\n",
      "BRAF --> BRAF --> PROPN --> NNP --> pobj --> XXXX --> True --> False\n",
      "( --> ( --> PUNCT --> -LRB- --> punct --> ( --> False --> False\n",
      "V600 --> V600 --> PROPN --> NNP --> subtok --> Xddd --> False --> False\n",
      ") --> ) --> PUNCT --> -RRB- --> punct --> ) --> False --> False\n",
      "mutation --> mutation --> NOUN --> NN --> npadvmod --> xxxx --> True --> False\n",
      "- --> - --> PUNCT --> HYPH --> punct --> - --> False --> False\n",
      "positive --> positive --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "unresectable --> unresectable --> ADJ --> JJ --> amod --> xxxx --> True --> False\n",
      "or --> or --> CCONJ --> CC --> cc --> xx --> True --> True\n",
      "metastatic --> metastatic --> ADJ --> JJ --> conj --> xxxx --> True --> False\n",
      "melanoma --> melanoma --> NOUN --> NN --> appos --> xxxx --> True --> False\n",
      ". --> . --> PUNCT --> . --> punct --> . --> False --> False\n",
      "\n",
      " --> \n",
      " --> SPACE --> _SP -->  --> \n",
      " --> False --> False\n",
      "\n",
      "#---- Dependency Parsing ----#\n",
      "\n",
      "The BRAF inhibitor dabrafenib --> dabrafenib --> nsubjpass --> indicated\n",
      "(Tafinlar(Â® --> ® --> appos --> dabrafenib\n",
      "the MEK inhibitor trametinib --> trametinib --> conj --> dabrafenib\n",
      "Mekinist(Â® --> ® --> appos --> trametinib\n",
      "combination --> combination --> pobj --> in\n",
      "the treatment --> treatment --> pobj --> for\n",
      "patients --> patients --> pobj --> of\n",
      "unresectable or metastatic melanoma --> melanoma --> pobj --> with\n",
      "a BRAF (V600) mutation --> mutation --> pobj --> with\n",
      "This article --> article --> nsubj --> reviews\n",
      "the therapeutic efficacy --> efficacy --> dobj --> reviews\n",
      "tolerability --> tolerability --> conj --> efficacy\n",
      "combination treatment --> treatment --> pobj --> of\n",
      "dabrafenib --> dabrafenib --> pobj --> with\n",
      "trametinib --> trametinib --> conj --> dabrafenib\n",
      "this indication --> indication --> pobj --> in\n",
      "relevant pharmacological data --> data --> dobj --> summarizes\n",
      "Dabrafenib --> Dabrafenib --> nsubj --> improved\n",
      "trametinib --> trametinib --> conj --> Dabrafenib\n",
      "significantly prolonged progression-free survival --> survival --> conj --> Dabrafenib\n",
      "PFS --> PFS --> appos --> survival\n",
      "overall survival --> survival --> conj --> survival\n",
      "OS --> OS --> appos --> survival\n",
      "objective response rates --> rates --> dobj --> improved\n",
      "ORRs --> ORRs --> appos --> rates\n",
      "preserved health-related quality --> quality --> conj --> rates\n",
      "life --> life --> pobj --> of\n",
      "HR-QOL --> QOL --> appos --> quality\n",
      "a greater extent --> extent --> pobj --> to\n",
      "dabrafenib --> dabrafenib --> pobj --> than\n",
      "the double-blind COMBI-d study --> study --> pobj --> in\n",
      "vemurafenib --> vemurafenib --> conj --> extent\n",
      "the open-label COMBI-v study --> study --> pobj --> in\n",
      "two large, randomized, phase III studies --> studies --> pobj --> in\n",
      "treatment --> treatment --> pobj --> in\n",
      "patients --> patients --> dobj --> improved\n",
      "unresectable or metastatic melanoma --> melanoma --> pobj --> with\n",
      "BRAF (V600E/K) mutation --> mutation --> pobj --> with\n",
      "Limited treatment benefit --> benefit --> nsubjpass --> seen\n",
      "the combination --> combination --> pobj --> with\n",
      "patients --> patients --> pobj --> in\n",
      "who --> who --> nsubj --> progressed\n",
      "prior BRAF inhibitor therapy --> therapy --> pobj --> on\n",
      "ORRs --> ORRs --> pobj --> by\n",
      "â‰¤ 15Â % --> % --> pobj --> of\n",
      "stable disease --> disease --> conj --> ORRs\n",
      "50Â % --> % --> dobj --> indicated\n",
      "patients --> patients --> pobj --> of\n",
      "small phase I and II studies --> studies --> pobj --> in\n",
      "Combination therapy --> therapy --> nsubj --> increase\n",
      "overall toxicity --> toxicity --> dobj --> increase\n",
      "dabrafenib --> dabrafenib --> pobj --> to\n",
      "vemurafenib --> vemurafenib --> conj --> dabrafenib\n",
      "most adverse events --> events --> pobj --> with\n",
      "AEs --> AEs --> appos --> events\n",
      "severity --> severity --> pobj --> in\n",
      "Fewer skin-related AEs --> AEs --> nsubjpass --> reported\n",
      "(e.g. cutaneous malignancies --> malignancies --> appos --> AEs\n",
      "hyperkeratinosis --> hyperkeratinosis --> conj --> malignancies\n",
      "hand-foot syndrome --> syndrome --> conj --> hyperkeratinosis\n",
      "combination therapy --> therapy --> pobj --> with\n",
      "dabrafenib --> dabrafenib --> pobj --> with\n",
      "vemurafenib --> vemurafenib --> conj --> dabrafenib\n",
      "reduced paradoxical activation --> activation --> pobj --> because\n",
      "the MAPK pathway --> pathway --> pobj --> of\n",
      "dabrafenib --> dabrafenib --> nsubj --> provides\n",
      "trametinib --> trametinib --> conj --> dabrafenib\n",
      "an important treatment option --> option --> dobj --> provides\n",
      "patients --> patients --> pobj --> for\n",
      "BRAF --> BRAF --> pobj --> with\n",
      "V600) mutation-positive unresectable or metastatic melanoma --> melanoma --> appos --> BRAF\n",
      "\n",
      "#---- Entity using Spacy ----#\n",
      "\n",
      "BRAF --> gene\n",
      "dabrafenib --> drug\n",
      "MEK --> protien\n",
      "trametinib --> drug\n",
      "melanoma --> diesease\n",
      "BRAF --> gene\n",
      "V600 --> varient\n",
      "trametinib --> drug\n",
      "Dabrafenib --> drug\n",
      "trametinib --> drug\n",
      "melanoma --> diesease\n",
      "BRAF --> gene\n",
      "V600E/K --> varient\n",
      "vemurafenib --> drug\n",
      "vemurafenib --> drug\n",
      "trametinib --> drug\n",
      "BRAF --> gene\n",
      "V600 --> varient\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "melanoma --> diesease\n"
     ]
    }
   ],
   "source": [
    "# POS tagging\n",
    "print('#---- POS Tagging ----#''\\n')\n",
    "for token in nlp_doc:\n",
    "    print(token.text,\"-->\", token.lemma_,\"-->\", token.pos_,\"-->\", token.tag_,\"-->\", token.dep_,\"-->\",\n",
    "            token.shape_,\"-->\", token.is_alpha,\"-->\", token.is_stop)\n",
    "\n",
    "# Dependency Parsing\n",
    "print('\\n'\"#---- Dependency Parsing ----#\"'\\n')\n",
    "for chunk in nlp_doc.noun_chunks:\n",
    "    print(chunk.text,\"-->\", chunk.root.text,\"-->\", chunk.root.dep_,\"-->\",\n",
    "            chunk.root.head.text)\n",
    "    \n",
    "# Entity Extration\n",
    "print('\\n'\"#---- Entity using Spacy ----#\"'\\n')\n",
    "for ent in nlp_doc.ents:\n",
    "    print(ent.text,\"-->\",ent.label_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tobias --> PERSON\n",
      "’m --> PERSON\n",
      "2018 --> DATE\n"
     ]
    }
   ],
   "source": [
    "nlp = spacy.load(\"en_core_web_lg\")\n",
    "for ent in nlp(doc1).ents:\n",
    "    print(ent.text, \"-->\", ent.label_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Playing with Displacy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [],
   "source": [
    "from spacy import displacy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "nlp = spacy.load(\"/home/ubuntu/Grakn/Spacy Play/spacy_en_core_web_lg\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [],
   "source": [
    "doc = nlp(TEXTS)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Custome NER Model Output using Displacy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div class=\"entities\" style=\"line-height: 2.5; direction: ltr\">The \n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    BRAF\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">gene</span>\n",
       "</mark>\n",
       " inhibitor \n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    dabrafenib\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">drug</span>\n",
       "</mark>\n",
       " (Tafinlar(Â®)) and the \n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    MEK\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">protien</span>\n",
       "</mark>\n",
       " inhibitor \n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    trametinib\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">drug</span>\n",
       "</mark>\n",
       " (Mekinist(Â®)) are indicated, as monotherapy or in combination with each other, for the treatment of patients with unresectable or metastatic \n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    melanoma\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">diesease</span>\n",
       "</mark>\n",
       " with a \n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    BRAF\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">gene</span>\n",
       "</mark>\n",
       " (\n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    V600\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">varient</span>\n",
       "</mark>\n",
       ") mutation. </br>This article reviews the therapeutic efficacy and tolerability of combination treatment with dabrafenib and \n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    trametinib\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">drug</span>\n",
       "</mark>\n",
       " in this indication and summarizes relevant pharmacological data. </br>\n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    Dabrafenib\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">drug</span>\n",
       "</mark>\n",
       " plus \n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    trametinib\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">drug</span>\n",
       "</mark>\n",
       " significantly prolonged progression-free survival (PFS) and overall survival (OS), improved objective response rates (ORRs) and preserved health-related quality of life (HR-QOL) to a greater extent than dabrafenib (in the double-blind COMBI-d study) and vemurafenib (in the open-label COMBI-v study) in two large, randomized, phase III studies in treatment-naÃ¯ve patients with unresectable or metastatic \n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    melanoma\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">diesease</span>\n",
       "</mark>\n",
       " with \n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    BRAF\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">gene</span>\n",
       "</mark>\n",
       " (\n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    V600E/K\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">varient</span>\n",
       "</mark>\n",
       ") mutation. Limited treatment benefit with the combination was also seen in patients who had progressed on prior BRAF inhibitor therapy, as indicated by ORRs of â‰¤ 15Â % and stable disease in â‰¤ 50Â % of patients in small phase I and II studies. Combination therapy did not increase overall toxicity relative to dabrafenib or \n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    vemurafenib\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">drug</span>\n",
       "</mark>\n",
       " monotherapy, with most adverse events (AEs) mild or moderate in severity and generally manageable. Fewer skin-related AEs (e.g. cutaneous malignancies, hyperkeratinosis and hand-foot syndrome) were reported with combination therapy than with dabrafenib or \n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    vemurafenib\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">drug</span>\n",
       "</mark>\n",
       ", probably because of reduced paradoxical activation of the MAPK pathway. Thus, dabrafenib plus \n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    trametinib\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">drug</span>\n",
       "</mark>\n",
       " provides an important treatment option for patients with \n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    BRAF\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">gene</span>\n",
       "</mark>\n",
       " (\n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    V600\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">varient</span>\n",
       "</mark>\n",
       ") mutation-positive unresectable or metastatic \n",
       "<mark class=\"entity\" style=\"background: #ddd; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    melanoma\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">diesease</span>\n",
       "</mark>\n",
       ".</div>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "displacy.render(doc, style=\"ent\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pretrained NER Output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [],
   "source": [
    "nlp = spacy.load(\"en_core_web_lg\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [],
   "source": [
    "doc = nlp(TEXTS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div class=\"entities\" style=\"line-height: 2.5; direction: ltr\">The \n",
       "<mark class=\"entity\" style=\"background: #7aecec; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    BRAF\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">ORG</span>\n",
       "</mark>\n",
       " inhibitor dabrafenib (Tafinlar(Â®)) and the \n",
       "<mark class=\"entity\" style=\"background: #7aecec; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    MEK\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">ORG</span>\n",
       "</mark>\n",
       " inhibitor trametinib (Mekinist(Â®)) are indicated, as monotherapy or in combination with each other, for the treatment of patients with unresectable or metastatic melanoma with a \n",
       "<mark class=\"entity\" style=\"background: #7aecec; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    BRAF\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">ORG</span>\n",
       "</mark>\n",
       " (V600) mutation. </br>This article reviews the therapeutic efficacy and tolerability of combination treatment with dabrafenib and trametinib in this indication and summarizes relevant pharmacological data. </br>Dabrafenib plus trametinib significantly prolonged progression-free survival (\n",
       "<mark class=\"entity\" style=\"background: #7aecec; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    PFS\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">ORG</span>\n",
       "</mark>\n",
       ") and overall survival (\n",
       "<mark class=\"entity\" style=\"background: #7aecec; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    OS\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">ORG</span>\n",
       "</mark>\n",
       "), improved objective response rates (ORRs) and preserved health-related quality of life (HR-QOL) to a greater extent than dabrafenib (in the double-blind COMBI-d study) and vemurafenib (in the open-label COMBI-v study) in \n",
       "<mark class=\"entity\" style=\"background: #e4e7d2; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    two\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">CARDINAL</span>\n",
       "</mark>\n",
       " large, randomized, phase III studies in treatment-naÃ¯ve patients with unresectable or metastatic melanoma with \n",
       "<mark class=\"entity\" style=\"background: #7aecec; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    BRAF\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">ORG</span>\n",
       "</mark>\n",
       " (V600E/K) mutation. Limited treatment benefit with the combination was also seen in patients who had progressed on prior \n",
       "<mark class=\"entity\" style=\"background: #7aecec; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    BRAF\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">ORG</span>\n",
       "</mark>\n",
       " inhibitor therapy, as indicated by ORRs of â‰¤ \n",
       "<mark class=\"entity\" style=\"background: #e4e7d2; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    15Â %\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">PERCENT</span>\n",
       "</mark>\n",
       " and stable disease in â‰¤ \n",
       "<mark class=\"entity\" style=\"background: #e4e7d2; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    50Â %\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">PERCENT</span>\n",
       "</mark>\n",
       " of patients in small phase I and II studies. Combination therapy did not increase overall toxicity relative to dabrafenib or vemurafenib monotherapy, with most adverse events (AEs) mild or moderate in severity and generally manageable. Fewer skin-related AEs (e.g. cutaneous malignancies, hyperkeratinosis and hand-foot syndrome) were reported with combination therapy than with dabrafenib or vemurafenib, probably because of reduced paradoxical activation of the \n",
       "<mark class=\"entity\" style=\"background: #7aecec; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    MAPK\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">ORG</span>\n",
       "</mark>\n",
       " pathway. Thus, dabrafenib plus trametinib provides an important treatment option for patients with \n",
       "<mark class=\"entity\" style=\"background: #7aecec; padding: 0.45em 0.6em; margin: 0 0.25em; line-height: 1; border-radius: 0.35em; box-decoration-break: clone; -webkit-box-decoration-break: clone\">\n",
       "    BRAF\n",
       "    <span style=\"font-size: 0.8em; font-weight: bold; line-height: 1; border-radius: 0.35em; text-transform: uppercase; vertical-align: middle; margin-left: 0.5rem\">ORG</span>\n",
       "</mark>\n",
       " (V600) mutation-positive unresectable or metastatic melanoma.</div>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "displacy.render(doc, style=\"ent\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Playing with Spacy - Entity Relation Extraction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "#!/usr/bin/env python\n",
    "# coding: utf8\n",
    "\"\"\"A simple example of extracting relations between phrases and entities using\n",
    "spaCy's named entity recognizer and the dependency parse. Here, we extract\n",
    "money and currency values (entities labelled as MONEY) and then check the\n",
    "dependency tree to find the noun phrase they are referring to – for example:\n",
    "$9.4 million --> Net income.\n",
    "\n",
    "Compatible with: spaCy v2.0.0+\n",
    "Last tested with: v2.1.0\n",
    "\"\"\"\n",
    "\n",
    "from __future__ import unicode_literals, print_function\n",
    "\n",
    "import plac\n",
    "import spacy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "TEXTS = [\n",
    "    \"Net income was $9.4 million compared to the prior year of $2.7 million.\",\n",
    "    \"Revenue exceeded twelve billion dollars, with a loss of $1b.\",\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [],
   "source": [
    "TEXTS = '''The BRAF inhibitor dabrafenib (Tafinlar(Â®)) and the MEK inhibitor trametinib (Mekinist(Â®)) are indicated, as monotherapy or in combination with each other, for the treatment of patients with unresectable or metastatic melanoma with a BRAF (V600) mutation. \\nThis article reviews the therapeutic efficacy and tolerability of combination treatment with dabrafenib and trametinib in this indication and summarizes relevant pharmacological data. \\nDabrafenib plus trametinib significantly prolonged progression-free survival (PFS) and overall survival (OS), improved objective response rates (ORRs) and preserved health-related quality of life (HR-QOL) to a greater extent than dabrafenib (in the double-blind COMBI-d study) and vemurafenib (in the open-label COMBI-v study) in two large, randomized, phase III studies in treatment-naÃ¯ve patients with unresectable or metastatic melanoma with BRAF (V600E/K) mutation. Limited treatment benefit with the combination was also seen in patients who had progressed on prior BRAF inhibitor therapy, as indicated by ORRs of â‰¤ 15Â % and stable disease in â‰¤ 50Â % of patients in small phase I and II studies. Combination therapy did not increase overall toxicity relative to dabrafenib or vemurafenib monotherapy, with most adverse events (AEs) mild or moderate in severity and generally manageable. Fewer skin-related AEs (e.g. cutaneous malignancies, hyperkeratinosis and hand-foot syndrome) were reported with combination therapy than with dabrafenib or vemurafenib, probably because of reduced paradoxical activation of the MAPK pathway. Thus, dabrafenib plus trametinib provides an important treatment option for patients with BRAF (V600) mutation-positive unresectable or metastatic melanoma.'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [],
   "source": [
    "@plac.annotations(model=(\"Model to load (needs parser and NER)\", \"positional\", None, str))\n",
    "\n",
    "def main(model=\"en_core_web_lg\"): ##  \"/home/ubuntu/Grakn/Spacy Play/spacy_en_core_web_lg\"\n",
    "    nlp = spacy.load(model)\n",
    "    print(\"Loaded model '%s'\" % model)\n",
    "    print(\"Processing %d texts\" % len(TEXTS))\n",
    "\n",
    "    for text in TEXTS:\n",
    "        doc = nlp(text)\n",
    "        relations = extract_currency_relations(doc)\n",
    "        for r1, r2 in relations:\n",
    "            print(\"{:<10}\\t{}\\t{}\".format(r1.text, r2.ent_type_, r2.text))\n",
    "\n",
    "\n",
    "def filter_spans(spans):\n",
    "    # Filter a sequence of spans so they don't contain overlaps\n",
    "    get_sort_key = lambda span: (span.end - span.start, span.start)\n",
    "    sorted_spans = sorted(spans, key=get_sort_key, reverse=True)\n",
    "    result = []\n",
    "    seen_tokens = set()\n",
    "    for span in sorted_spans:\n",
    "        if span.start not in seen_tokens and span.end - 1 not in seen_tokens:\n",
    "            result.append(span)\n",
    "            seen_tokens.update(range(span.start, span.end))\n",
    "    return result\n",
    "\n",
    "\n",
    "def extract_currency_relations(doc):\n",
    "    # Merge entities and noun chunks into one token\n",
    "    seen_tokens = set()\n",
    "    spans = list(doc.ents) + list(doc.noun_chunks)\n",
    "    spans = filter_spans(spans)\n",
    "    with doc.retokenize() as retokenizer:\n",
    "        for span in spans:\n",
    "            retokenizer.merge(span)\n",
    "\n",
    "    relations = []\n",
    "    for money in filter(lambda w: w.ent_type_ ==\"MONEY\" , doc): ##   \"gene\"                      \n",
    "        if money.dep_ in (\"attr\", \"dobj\"):\n",
    "            subject = [w for w in money.head.lefts if w.dep_ == \"nsubj\"]\n",
    "            if subject:\n",
    "                subject = subject[0]\n",
    "                relations.append((subject, money))\n",
    "        elif money.dep_ == \"pobj\" and money.head.dep_ == \"prep\":\n",
    "            relations.append((money.head.head, money))\n",
    "    return relations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "main()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "if __name__ == \"__main__\":\n",
    "    plac.call(main)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}