ed14c8b2-2ac1-41e0-acea-3cc55cd94f83.txt

import os
import sys
with open(sys.argv[0]) as f:
    code = f.read() # read the code of this file ASAP, for logging
import uuid
import glob
import time
import contextlib
from dataclasses import dataclass

import numpy as np
import torch
from torch import nn
import torch.nn.functional as F
import torch.distributed as dist
import torch._inductor.config as config
from torch.nn.parallel import DistributedDataParallel as DDP
# Use of FlexAttention contributed by @KoszarskyB
from torch.nn.attention.flex_attention import flex_attention, create_block_mask
flex_attention = torch.compile(flex_attention, dynamic=False)
create_block_mask = torch.compile(create_block_mask, dynamic=False)

# -----------------------------------------------------------------------------
# Muon optimizer

def zeropower_via_svd(G, steps=None):
    U, S, V = G.svd()
    return U @ V.T

@torch.compile
def zeropower_via_newtonschulz5(G, steps=10, eps=1e-7):
    """
    Newton-Schulz iteration to compute the zeroth power / orthogonalization of G. We opt to use a
    quintic iteration whose coefficients are selected to maximize the slope at zero. For the purpose
    of minimizing steps, it turns out to be empirically effective to keep increasing the slope at
    zero even beyond the point where the iteration no longer converges all the way to one everywhere
    on the interval. This iteration therefore does not produce UV^T but rather something like US'V^T
    where S' is diagonal with S_{ii}' ~ Uniform(0.5, 1.5), which turns out not to hurt model
    performance at all relative to UV^T, where USV^T = G is the SVD.
    """
    assert len(G.shape) == 2
    a, b, c = (3.4445, -4.7750,  2.0315)
    X = G.bfloat16()
    X /= (X.norm() + eps) # ensure top singular value <= 1
    if G.size(0) > G.size(1):
        X = X.T
    for _ in range(steps):
        A = X @ X.T
        B = b * A + c * A @ A # adapted from suggestion by @jxbz, @leloykun, and @YouJiacheng
        X = a * X + B @ X
    if G.size(0) > G.size(1):
        X = X.T
    return X

zeropower_backends = dict(svd=zeropower_via_svd, newtonschulz5=zeropower_via_newtonschulz5)

class Muon(torch.optim.Optimizer):
    """
    Muon - MomentUm Orthogonalized by Newton-schulz

    Muon internally runs standard SGD-momentum, and then performs an orthogonalization post-
    processing step, in which each 2D parameter's update is replaced with the nearest orthogonal
    matrix. To efficiently orthogonalize each update, we use a Newton-Schulz iteration, which has
    the advantage that it can be stably run in bfloat16 on the GPU.

    Some warnings:
    - This optimizer assumes that all parameters passed in are 2D.
    - It should not be used for the embedding layer, the final fully connected layer, or any {0,1}-D
    parameters; those should all be optimized by a standard method (e.g., AdamW).
    - To use it with 4D convolutional filters, it works well to just flatten their last 3 dimensions.
    - We believe it is unlikely to work well for training with small batch size.
    - We believe it may not work well for finetuning pretrained models, but we haven't tested this.
    - We have not yet tried this optimizer for training scenarios larger than NanoGPT (124M).

    Arguments:
        lr: The learning rate used by the internal SGD.
        momentum: The momentum used by the internal SGD.
        nesterov: Whether to use Nesterov-style momentum in the internal SGD. (recommended)
        backend: The chosen backend for the orthogonalization step. (recommended: 'newtonschulz5')
        backend_steps: The number of iteration steps to use in the backend, if it is iterative.
    """
    def __init__(self, params, lr=0.02, momentum=0.95, nesterov=True,
                 backend='newtonschulz5', backend_steps=5):
        defaults = dict(lr=lr, momentum=momentum, nesterov=nesterov, backend=backend, backend_steps=backend_steps)
        super().__init__(params, defaults)

    def step(self):

        for group in self.param_groups:

            lr = group['lr']
            momentum = group['momentum']
            zeropower_backend = zeropower_backends[group['backend']]

            # generate weight updates in distributed fashion
            total_params = sum(p.numel() for p in group['params'])
            updates_flat = torch.zeros(total_params, device='cuda', dtype=torch.bfloat16)
            curr_idx = 0
            for i, p in enumerate(group['params']):
                # luckily this will perfectly distribute a transformer with multiple of 4 layers to 8 GPUs
                if i % int(os.environ['WORLD_SIZE']) == int(os.environ['RANK']):
                    g = p.grad
                    assert g is not None
                    state = self.state[p]
                    if 'momentum_buffer' not in state:
                        state['momentum_buffer'] = torch.zeros_like(g)
                    buf = state['momentum_buffer']
                    buf.mul_(momentum).add_(g)
                    g = g.add(buf, alpha=momentum) if group['nesterov'] else buf
                    g = zeropower_backend(g, steps=group['backend_steps'])
                    g *= max(1, g.size(0)/g.size(1))**0.5
                    updates_flat[curr_idx:curr_idx+p.numel()] = g.flatten()
                curr_idx += p.numel()

            # sync updates across devices. we are not memory-constrained so can do this simple deserialization
            dist.all_reduce(updates_flat, op=dist.ReduceOp.SUM)

            # deserialize and apply updates
            curr_idx = 0
            for p in group['params']:
                g = updates_flat[curr_idx:curr_idx+p.numel()].view_as(p.data).type_as(p.data)
                p.data.add_(g, alpha=-lr)
                curr_idx += p.numel()

# -----------------------------------------------------------------------------
# PyTorch nn.Module definitions for the GPT-2 model

def norm(x):
    return F.rms_norm(x, (x.size(-1),))

class CastedLinear(nn.Linear):

    def __init__(self, in_features, out_features):
        super().__init__(in_features, out_features, bias=False)

    def forward(self, x):
        return F.linear(x, self.weight.to(x.dtype))

class Rotary(torch.nn.Module):

    def __init__(self, dim, base=10000):
        super().__init__()
        self.register_buffer('inv_freq', (1 / base) ** (torch.arange(0, dim, 2) / dim))
        self.seq_len_cached = None
        self.cos_cached = None
        self.sin_cached = None

    def forward(self, x):
        seq_len = x.shape[1]
        if seq_len != self.seq_len_cached:
            t = torch.arange(seq_len, device=x.device)
            freqs = torch.outer(t, self.inv_freq)
            self.seq_len_cached = seq_len
            self.cos_cached = freqs.cos()
            self.sin_cached = freqs.sin()
        cos, sin = self.cos_cached[None, :, None, :], self.sin_cached[None, :, None, :]
        # apply_rotary_emb(x, cos, sin)
        x1, x2 = x.chunk(2, dim=3)
        y1 = x1 * cos + x2 * sin
        y2 = x1 * (-sin) + x2 * cos
        return torch.cat((y1, y2), 3).type_as(x)

class CausalSelfAttention(nn.Module):

    def __init__(self, dim, n_head):
        super().__init__()
        assert dim % n_head == 0
        self.n_head = n_head
        self.c_q = CastedLinear(dim, dim)
        self.c_k = CastedLinear(dim, dim)
        self.c_v = CastedLinear(dim, dim)
        # value residual lambda
        self.lamb = nn.Parameter(torch.tensor(0.5)) # @Grad62304977
        # rotary embeddings
        self.rotary = Rotary(dim // n_head) # dim // n_head = head_dim
        # output projection
        self.c_proj = CastedLinear(dim, dim)
        self.c_proj.weight.data.zero_() # zero init suggested by @Grad62304977

    def forward(self, x, vi, block_mask):
        B, T = x.size(0), x.size(1) # batch size, sequence length
        assert B == 1, "Must use batch size = 1 for FlexAttention"
        q = self.c_q(x).view(B, T, self.n_head, -1)
        k = self.c_k(x).view(B, T, self.n_head, -1)
        v = self.c_v(x).view(B, T, self.n_head, -1)
        v = (1 - self.lamb) * v + self.lamb * vi.view_as(v) # @Grad62304977
        q, k = norm(q), norm(k) # QK norm suggested by @Grad62304977
        q, k = self.rotary(q), self.rotary(k)
        y = flex_attention(q.transpose(1, 2), k.transpose(1, 2), v.transpose(1, 2), block_mask=block_mask)
        y = y.transpose(1, 2).contiguous().view_as(x) # re-assemble all head outputs side by side
        y = self.c_proj(y)
        return y

class MLP(nn.Module):

    def __init__(self, dim):
        super().__init__()
        self.c_fc   = CastedLinear(dim, 4 * dim)
        self.c_proj = CastedLinear(4 * dim, dim)
        self.c_proj.weight.data.zero_() # zero init suggested by @Grad62304977

    def forward(self, x):
        x = self.c_fc(x)
        x = F.relu(x).square() # https://arxiv.org/abs/2109.08668v2; ~1-2% better than GELU; suggested by @SKYLINEZ007 and @Grad62304977
        x = self.c_proj(x)
        return x

class Block(nn.Module):

    def __init__(self, config):
        super().__init__()
        self.attn = CausalSelfAttention(config.n_embd, config.n_head)
        self.mlp = MLP(config.n_embd)
        self.lambdas = nn.Parameter(torch.tensor([1., 0.]))

    def forward(self, x, vi, x0, block_mask):
        x = self.lambdas[0] * x + self.lambdas[1] * x0
        x = x + self.attn(norm(x), vi, block_mask)
        x = x + self.mlp(norm(x))
        return x

# -----------------------------------------------------------------------------
# The main GPT-2 model

@dataclass
class GPTConfig:
    vocab_size : int = 50304
    n_layer : int = 12
    n_head : int = 6 # head dim 128 suggested by @Grad62304977
    n_embd : int = 768

class GPT(nn.Module):

    def __init__(self, config):
        super().__init__()

        # U-net design by @brendanh0gan
        self.num_encoder_layers = config.n_layer // 2 # Half of the layers for encoder
        self.num_decoder_layers = config.n_layer - self.num_encoder_layers # Remaining for decoder
        # Add learnable skip connection weights for decoder layers
        self.skip_weights = nn.Parameter(torch.ones(self.num_decoder_layers))

        self.transformer = nn.ModuleDict(dict(
            wte = nn.Embedding(config.vocab_size, config.n_embd),
            # token value embeddings by @KoszarskyB - inspired by @Grad62304977's value residual learning
            vte = nn.Embedding(config.vocab_size, config.n_embd*12),
            h = nn.ModuleList([Block(config) for _ in range(config.n_layer)]),
        ))
        self.lm_head = CastedLinear(config.n_embd, config.vocab_size)
        self.lm_head.weight.data.zero_() # @Grad62304977

    def forward(self, idx, target, attn_blocksize):

        docs = (idx == 50256).cumsum(0)
        def document_causal_mask(b, h, q_idx, kv_idx):
          causal_mask = q_idx >= kv_idx
          document_mask = docs[q_idx] == docs[kv_idx]
          window_mask = q_idx - kv_idx < attn_blocksize
          return causal_mask & document_mask & window_mask

        S = len(idx)
        block_mask = create_block_mask(document_causal_mask, None, None, S, S, device="cuda", _compile=True)

        # forward the GPT model itself
        x = self.transformer.wte(idx[None]) # token embeddings of shape (b, t, n_embd)
        x = norm(x) # @Grad62304977
        x0 = x
        vi = self.transformer.vte(idx[None]).chunk(12, dim=-1)

        # Store outputs for U-Net skip connections
        skip_connections = []
        # Encoder pass - process only the first half of the blocks
        for i in range(self.num_encoder_layers):
            x = self.transformer.h[i](x, vi[i], x0, block_mask)
            skip_connections.append(x)
        # Decoder pass - process the remaining blocks with weighted skip connections
        for i in range(self.num_decoder_layers):
            x = x + self.skip_weights[i] * skip_connections.pop()
            x = self.transformer.h[self.num_encoder_layers + i](x, vi[self.num_encoder_layers+i], x0, block_mask)

        x = norm(x)
        logits = self.lm_head(x)
        logits = 30 * torch.tanh(logits / 30) # @Grad62304977
        logits = logits.float()
        loss = F.cross_entropy(logits.view(-1, logits.size(-1)), target.view(-1))
        return loss

# -----------------------------------------------------------------------------
# Our own simple Distributed Data Loader

def _peek_data_shard(filename):
    # only reads the header, returns header data
    with open(filename, "rb") as f:
        # first read the header, which is 256 int32 integers (4 bytes each)
        header = np.frombuffer(f.read(256*4), dtype=np.int32)
    if header[0] != 20240520:
        print("ERROR: magic number mismatch in the data .bin file!")
        print("---> HINT: Are you passing in a correct file with --input_bin?")
        print("---> HINT: Dataset encoding changed recently, re-run data prepro or refer again to README")
        print("---> HINT: For example re-run: `python dev/data/tinyshakespeare.py`, then re-try")
        exit(1)
    assert header[1] == 1, "unsupported version"
    ntok = header[2] # number of tokens (claimed)
    return ntok # for now just return the number of tokens

def _load_data_shard(filename):
    with open(filename, "rb") as f:
        # first read the header, which is 256 int32 integers (4 bytes each)
        header = np.frombuffer(f.read(256*4), dtype=np.int32)
        assert header[0] == 20240520, "magic number mismatch in the data .bin file"
        assert header[1] == 1, "unsupported version"
        ntok = header[2] # number of tokens (claimed)
        # the rest of it are tokens, stored as uint16
        tokens = np.frombuffer(f.read(), dtype=np.uint16)
    assert len(tokens) == ntok, "number of tokens read does not match header?"
    return tokens

class DistributedDataLoader:
    def __init__(self, filename_pattern, T, process_rank, num_processes):
        self.process_rank = process_rank
        self.num_processes = num_processes
        self.T = T

        # glob files that match the pattern
        self.files = sorted(glob.glob(filename_pattern))
        assert len(self.files) > 0, f"did not find any files that match the pattern {filename_pattern}"

        # load and validate all data shards, count number of tokens in total
        ntok_total = 0
        for fname in self.files:
            shard_ntok = _peek_data_shard(fname)
            assert shard_ntok >= num_processes * T + 1
            ntok_total += int(shard_ntok)
        self.ntok_total = ntok_total

        self.reset()

    def reset(self):
        self.current_shard = -1
        self.advance()

    def advance(self): # advance to next data shard
        self.current_shard = (self.current_shard + 1) % len(self.files)
        self.current_position = self.process_rank * self.T
        self.tokens = _load_data_shard(self.files[self.current_shard])

    def next_batch(self):
        batch_size = self.T * self.num_processes
        buf = self.tokens[self.current_position:self.current_position+self.T+1]
        buf = torch.tensor(buf.astype(np.int32), dtype=torch.long)
        x = buf[:-1] # inputs
        y = buf[1:] # targets
        # advance current position and load next shard if necessary
        self.current_position += batch_size
        if self.current_position + batch_size >= len(self.tokens):
            self.advance()
        return x.cuda(), y.cuda()

# -----------------------------------------------------------------------------
# int main

@dataclass
class Hyperparameters:
    # data hyperparams
    input_bin : str = 'data/fineweb10B/fineweb_train_*.bin' # input .bin to train on
    input_val_bin : str = 'data/fineweb10B/fineweb_val_*.bin' # input .bin to eval validation loss on
    # optimization hyperparams
    batch_size : int = 8 # batch size, in sequences, across all devices
    sequence_length : int = 64*1024 # sequence length, in tokens
    num_iterations : int = 1530 # number of iterations to run
    warmup_iters : int = 0
    cooldown_iters : int = 600 # number of iterations of linear warmup/cooldown for triangular or trapezoidal schedule
    weight_decay : float = 0
    # evaluation and logging hyperparams
    val_loss_every : int = 125 # every how many steps to evaluate val loss? 0 for only at the end
    val_tokens : int = 10485760 # how many tokens of validation data? it's important to keep this fixed for consistent comparisons
    save_every : int = 0 # every how many steps to save the checkpoint? 0 for only at the end
args = Hyperparameters()

# set up DDP (distributed data parallel). torchrun sets this env variable
assert torch.cuda.is_available()
dist.init_process_group(backend='nccl')
ddp_rank = int(os.environ['RANK'])
ddp_local_rank = int(os.environ['LOCAL_RANK'])
ddp_world_size = int(os.environ['WORLD_SIZE'])
device = f'cuda:{ddp_local_rank}'
torch.cuda.set_device(device)
print(f"using device: {device}")
master_process = (ddp_rank == 0) # this process will do logging, checkpointing etc.

# begin logging
logfile = None
if master_process:
    run_id = str(uuid.uuid4())
    logdir = 'logs/%s/' % run_id
    os.makedirs(logdir, exist_ok=True)
    logfile = 'logs/%s.txt' % run_id
    # create the log file
    with open(logfile, "w") as f:
        # begin the log by printing this file (the Python code)
        f.write(code)
        f.write('='*100 + '\n')
def print0(s, logonly=False):
    if master_process:
        with open(logfile, "a") as f:
            if not logonly:
                print(s)
            f.write(s+'\n')
# log information about the hardware/software environment this is running on
# and print the full `nvidia-smi` to file
print0(f"Running pytorch {torch.version.__version__} compiled for CUDA {torch.version.cuda}\nnvidia-smi:")
import subprocess
result = subprocess.run(['nvidia-smi'], stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True)
print0(f'{result.stdout}', logonly=True)
print0('='*100, logonly=True)

# convenience variables
T = args.sequence_length
# calculate the number of steps to take in the val loop.
assert args.val_tokens % (T * ddp_world_size) == 0
val_steps = args.val_tokens // (T * ddp_world_size)
# calculate the steps of gradient accumulation required to attain the desired global batch size.
assert args.batch_size % (ddp_world_size) == 0
train_accumulation_steps = args.batch_size // ddp_world_size

# load tokens
train_loader = DistributedDataLoader(args.input_bin, T, ddp_rank, ddp_world_size)
val_loader = DistributedDataLoader(args.input_val_bin, T, ddp_rank, ddp_world_size)
print0(f"Training DataLoader: total number of tokens: {train_loader.ntok_total} across {len(train_loader.files)} files")
print0(f"Validation DataLoader: total number of tokens: {val_loader.ntok_total} across {len(val_loader.files)} files")
print0('='*100, logonly=True)
x, y = train_loader.next_batch()

# there are only 50257 unique GPT-2 tokens; we extend to nearest multiple of 128 for efficiency. suggested to me by @Grad62304977.
# this originates from Karpathy's experiments.
num_vocab = 50304
model = GPT(GPTConfig(vocab_size=num_vocab, n_layer=12, n_head=6, n_embd=768))
model = model.cuda().bfloat16()
for m in model.modules():
    if isinstance(m, CastedLinear):
        m.float()
if hasattr(config, "coordinate_descent_tuning"):
    config.coordinate_descent_tuning = True # suggested by @Chillee
model = torch.compile(model)
# here we wrap model into DDP container
model = DDP(model, device_ids=[ddp_local_rank])
raw_model = model.module # always contains the "raw" unwrapped model

# init the optimizer(s)
optimizer1 = torch.optim.Adam([raw_model.transformer.wte.weight, raw_model.transformer.vte.weight], lr=0.6, betas=(0.8, 0.95), fused=True)
optimizer2 = torch.optim.Adam([raw_model.lm_head.weight], lr=0.008, betas=(0.8, 0.95), fused=True)
params = list(raw_model.transformer.h.parameters())
matrix_params = [p for p in params if p.ndim == 2]
scalar_params = [p for p in params if p.ndim < 2] + [raw_model.skip_weights]
optimizer3 = Muon(matrix_params, lr=0.05, momentum=0.95)
optimizer4 = torch.optim.Adam(scalar_params, lr=0.04, betas=(0.8, 0.95), fused=True) # note that this learning rate is neither sensitive nor tuned
optimizers = [optimizer1, optimizer2, optimizer3, optimizer4]
# learning rate decay scheduler (linear warmup and cooldown)
def get_lr(it):
    assert it <= args.num_iterations
    # 1) linear warmup for warmup_iters steps
    if it < args.warmup_iters:
        return (it+1) / args.warmup_iters
    # 2) constant lr for a while
    elif it < args.num_iterations - args.cooldown_iters:
        return 1.0
    # 3) linear cooldown
    else:
        decay_ratio = (args.num_iterations - it) / args.cooldown_iters
        return decay_ratio
schedulers = [torch.optim.lr_scheduler.LambdaLR(opt, get_lr) for opt in optimizers]

# Start training loop
training_time_ms = 0
# start the clock
torch.cuda.synchronize()
t0 = time.time()
# begin training
for step in range(args.num_iterations + 1):
    last_step = (step == args.num_iterations)
    # This effectively ignores timing first 10 steps, which are slower for weird reasons.
    # Alternately, and slightly more correctly in terms of benchmarking, we could do 10
    # steps with dummy data first, and then re-initialize the model and reset the loader.
    if step == 10:
        training_time_ms = 0
        t0 = time.time()
    timed_steps = float('nan') if step <= 11 else (step - 10) + 1 # <= 11 to avoid bug in val

    # Set the attention blocksize for the current step, in chunks of 64. By @fernbear.bsky.social
    attn_blocksize = torch.tensor(64*((step/args.num_iterations * (1792 - 64) + 64)//64), dtype=torch.int, device='cuda')

    # once in a while evaluate the validation dataset
    if (last_step or (args.val_loss_every > 0 and step % args.val_loss_every == 0)):
        # stop the clock
        torch.cuda.synchronize()
        training_time_ms += 1000 * (time.time() - t0)
        # run validation batches
        model.eval()
        val_loader.reset()
        val_loss = 0.0
        for _ in range(val_steps):
            with torch.no_grad():
                x_val, y_val = val_loader.next_batch()
                val_loss += model(x_val, y_val, attn_blocksize=attn_blocksize)
        dist.all_reduce(val_loss, op=dist.ReduceOp.AVG)
        val_loss /= val_steps
        # log val loss to console and to logfile
        print0(f'step:{step}/{args.num_iterations} val_loss:{val_loss:.4f} train_time:{training_time_ms:.0f}ms step_avg:{training_time_ms/(timed_steps-1):.2f}ms')
        # start the clock again
        torch.cuda.synchronize()
        t0 = time.time()

    if master_process and (last_step or (args.save_every > 0 and step % args.save_every == 0)):
        # stop the clock
        torch.cuda.synchronize()
        training_time_ms += 1000 * (time.time() - t0)
        # save the state of the training process
        log = dict(step=step, code=code, model=raw_model.state_dict(), optimizers=[opt.state_dict() for opt in optimizers])
        torch.save(log, 'logs/%s/state_step%06d.pt' % (run_id, step))
        # start the clock again
        torch.cuda.synchronize()
        t0 = time.time()

    # bit confusing: we want to make sure to eval on 0th iteration
    # but also after the very last iteration. so we loop for step <= num_iterations
    # instead of just < num_iterations (one extra due to <=), only to do
    # the validation/sampling one last time, and then we break right here as we're done.
    if last_step:
        break

    # --------------- TRAINING SECTION BEGIN -----------------
    model.train()
    for i in range(1, train_accumulation_steps+1):
        ctx = model.no_sync() if i < train_accumulation_steps else contextlib.nullcontext()
        with ctx: # there's no need to sync gradients every accumulation step
            # forward pass
            loss = model(x, y, attn_blocksize=attn_blocksize)
            # advance the dataset for the next batch
            x, y = train_loader.next_batch()
            # backward pass
            loss.backward()
        train_loss = loss.detach()
    for p in model.parameters():
        p.grad /= train_accumulation_steps
    # momentum warmup for Muon
    frac = min(step/300, 1)
    optimizer3.param_groups[0]['momentum'] = (1 - frac) * 0.85 + frac * 0.95
    # step the optimizers and schedulers
    for opt, sched in zip(optimizers, schedulers):
        opt.step()
        sched.step()
    # null the gradients
    model.zero_grad(set_to_none=True)
    # --------------- TRAINING SECTION END -------------------
    # everything that follows now is just diagnostics, prints, logging, etc.

    #dist.all_reduce(train_loss, op=dist.ReduceOp.AVG) # all-reducing the training loss would be more correct in terms of logging, but slower
    approx_time = training_time_ms + 1000 * (time.time() - t0)
    print0(f"step:{step+1}/{args.num_iterations} train_loss:{train_loss.item():.4f} train_time:{approx_time:.0f}ms step_avg:{approx_time/timed_steps:.2f}ms")

if master_process:
    print(f"peak memory consumption: {torch.cuda.max_memory_allocated() // 1024 // 1024} MiB")

# -------------------------------------------------------------------------
# clean up nice
dist.destroy_process_group()
====================================================================================================
Running pytorch 2.6.0.dev20241203+cu124 compiled for CUDA 12.4
nvidia-smi:
Thu Dec  5 03:41:24 2024       
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 535.183.06             Driver Version: 535.183.06   CUDA Version: 12.2     |
|-----------------------------------------+----------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |         Memory-Usage | GPU-Util  Compute M. |
|                                         |                      |               MIG M. |
|=========================================+======================+======================|
|   0  NVIDIA H100 80GB HBM3          On  | 00000000:19:00.0 Off |                    0 |
| N/A   39C    P0              75W / 700W |      3MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   1  NVIDIA H100 80GB HBM3          On  | 00000000:3B:00.0 Off |                    0 |
| N/A   31C    P0             115W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   2  NVIDIA H100 80GB HBM3          On  | 00000000:4C:00.0 Off |                    0 |
| N/A   31C    P0             118W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   3  NVIDIA H100 80GB HBM3          On  | 00000000:5D:00.0 Off |                    0 |
| N/A   38C    P0             119W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   4  NVIDIA H100 80GB HBM3          On  | 00000000:9B:00.0 Off |                    0 |
| N/A   39C    P0             123W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   5  NVIDIA H100 80GB HBM3          On  | 00000000:BB:00.0 Off |                    0 |
| N/A   30C    P0             110W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   6  NVIDIA H100 80GB HBM3          On  | 00000000:CB:00.0 Off |                    0 |
| N/A   39C    P0             127W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   7  NVIDIA H100 80GB HBM3          On  | 00000000:DB:00.0 Off |                    0 |
| N/A   30C    P0             119W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
                                                                                         
+---------------------------------------------------------------------------------------+
| Processes:                                                                            |
|  GPU   GI   CI        PID   Type   Process name                            GPU Memory |
|        ID   ID                                                             Usage      |
|=======================================================================================|
+---------------------------------------------------------------------------------------+

====================================================================================================
Training DataLoader: total number of tokens: 1100000000 across 11 files
Validation DataLoader: total number of tokens: 100000000 across 1 files
====================================================================================================
step:0/1530 val_loss:10.8258 train_time:0ms step_avg:nanms
step:1/1530 train_loss:10.8258 train_time:31472ms step_avg:nanms
step:2/1530 train_loss:10.0847 train_time:31583ms step_avg:nanms
step:3/1530 train_loss:8.4120 train_time:31743ms step_avg:nanms
step:4/1530 train_loss:7.5646 train_time:31903ms step_avg:nanms
step:5/1530 train_loss:7.4415 train_time:32064ms step_avg:nanms
step:6/1530 train_loss:6.9575 train_time:32224ms step_avg:nanms
step:7/1530 train_loss:7.1982 train_time:32384ms step_avg:nanms
step:8/1530 train_loss:6.7300 train_time:32546ms step_avg:nanms
step:9/1530 train_loss:6.6212 train_time:32705ms step_avg:nanms
step:10/1530 train_loss:6.5043 train_time:32865ms step_avg:nanms
step:11/1530 train_loss:6.4671 train_time:114ms step_avg:nanms
step:12/1530 train_loss:6.3423 train_time:274ms step_avg:nanms
step:13/1530 train_loss:6.2625 train_time:434ms step_avg:144.65ms
step:14/1530 train_loss:6.1986 train_time:593ms step_avg:148.31ms
step:15/1530 train_loss:6.1737 train_time:753ms step_avg:150.52ms
step:16/1530 train_loss:6.0992 train_time:914ms step_avg:152.32ms
step:17/1530 train_loss:6.1649 train_time:1075ms step_avg:153.51ms
step:18/1530 train_loss:5.9431 train_time:1235ms step_avg:154.32ms
step:19/1530 train_loss:6.0127 train_time:1395ms step_avg:155.05ms
step:20/1530 train_loss:5.6685 train_time:1556ms step_avg:155.57ms
step:21/1530 train_loss:5.9396 train_time:1717ms step_avg:156.07ms
step:22/1530 train_loss:6.1910 train_time:1876ms step_avg:156.37ms
step:23/1530 train_loss:5.8741 train_time:2038ms step_avg:156.75ms
step:24/1530 train_loss:6.0123 train_time:2198ms step_avg:156.97ms
step:25/1530 train_loss:5.6859 train_time:2356ms step_avg:157.10ms
step:26/1530 train_loss:5.5911 train_time:2518ms step_avg:157.40ms
step:27/1530 train_loss:5.7918 train_time:2680ms step_avg:157.66ms
step:28/1530 train_loss:5.3947 train_time:2840ms step_avg:157.78ms
step:29/1530 train_loss:5.6829 train_time:3001ms step_avg:157.97ms
step:30/1530 train_loss:5.4682 train_time:3162ms step_avg:158.09ms
step:31/1530 train_loss:5.4392 train_time:3322ms step_avg:158.20ms
step:32/1530 train_loss:5.2740 train_time:3482ms step_avg:158.25ms
step:33/1530 train_loss:5.5741 train_time:3642ms step_avg:158.35ms
step:34/1530 train_loss:5.4848 train_time:3802ms step_avg:158.40ms
step:35/1530 train_loss:5.6208 train_time:3962ms step_avg:158.50ms
step:36/1530 train_loss:5.5445 train_time:4123ms step_avg:158.60ms
step:37/1530 train_loss:5.4549 train_time:4283ms step_avg:158.62ms
step:38/1530 train_loss:5.2936 train_time:4444ms step_avg:158.70ms
step:39/1530 train_loss:5.3191 train_time:4604ms step_avg:158.75ms
step:40/1530 train_loss:5.2258 train_time:4764ms step_avg:158.80ms
step:41/1530 train_loss:5.2244 train_time:4925ms step_avg:158.86ms
step:42/1530 train_loss:5.1653 train_time:5084ms step_avg:158.87ms
step:43/1530 train_loss:5.2563 train_time:5244ms step_avg:158.92ms
step:44/1530 train_loss:5.2355 train_time:5405ms step_avg:158.96ms
step:45/1530 train_loss:5.3789 train_time:5565ms step_avg:159.00ms
step:46/1530 train_loss:5.1873 train_time:5725ms step_avg:159.04ms
step:47/1530 train_loss:5.0905 train_time:5885ms step_avg:159.04ms
step:48/1530 train_loss:5.2136 train_time:6045ms step_avg:159.09ms
step:49/1530 train_loss:5.1603 train_time:6205ms step_avg:159.11ms
step:50/1530 train_loss:5.2518 train_time:6366ms step_avg:159.15ms
step:51/1530 train_loss:5.1337 train_time:6526ms step_avg:159.18ms
step:52/1530 train_loss:5.0206 train_time:6686ms step_avg:159.20ms
step:53/1530 train_loss:5.1746 train_time:6847ms step_avg:159.24ms
step:54/1530 train_loss:5.0075 train_time:7008ms step_avg:159.27ms
step:55/1530 train_loss:5.4065 train_time:7169ms step_avg:159.30ms
step:56/1530 train_loss:5.0048 train_time:7329ms step_avg:159.32ms
step:57/1530 train_loss:4.8780 train_time:7490ms step_avg:159.36ms
step:58/1530 train_loss:5.0496 train_time:7651ms step_avg:159.39ms
step:59/1530 train_loss:5.0372 train_time:7811ms step_avg:159.40ms
step:60/1530 train_loss:5.1790 train_time:7972ms step_avg:159.44ms
step:61/1530 train_loss:4.8739 train_time:8133ms step_avg:159.48ms
step:62/1530 train_loss:4.9735 train_time:8293ms step_avg:159.48ms
step:63/1530 train_loss:4.9723 train_time:8453ms step_avg:159.50ms
step:64/1530 train_loss:4.9348 train_time:8614ms step_avg:159.52ms
step:65/1530 train_loss:4.7893 train_time:8774ms step_avg:159.53ms
step:66/1530 train_loss:4.8957 train_time:8936ms step_avg:159.57ms
step:67/1530 train_loss:4.8038 train_time:9096ms step_avg:159.57ms
step:68/1530 train_loss:5.0879 train_time:9255ms step_avg:159.58ms
step:69/1530 train_loss:4.7200 train_time:9415ms step_avg:159.57ms
step:70/1530 train_loss:4.8256 train_time:9575ms step_avg:159.58ms
step:71/1530 train_loss:4.9615 train_time:9736ms step_avg:159.61ms
step:72/1530 train_loss:4.8625 train_time:9897ms step_avg:159.63ms
step:73/1530 train_loss:4.7636 train_time:10057ms step_avg:159.63ms
step:74/1530 train_loss:4.9036 train_time:10219ms step_avg:159.67ms
step:75/1530 train_loss:4.8196 train_time:10379ms step_avg:159.67ms
step:76/1530 train_loss:4.7850 train_time:10539ms step_avg:159.69ms
step:77/1530 train_loss:4.9021 train_time:10701ms step_avg:159.72ms
step:78/1530 train_loss:5.1422 train_time:10861ms step_avg:159.73ms
step:79/1530 train_loss:4.8201 train_time:11023ms step_avg:159.75ms
step:80/1530 train_loss:4.8538 train_time:11183ms step_avg:159.76ms
step:81/1530 train_loss:4.6345 train_time:11343ms step_avg:159.77ms
step:82/1530 train_loss:4.8062 train_time:11504ms step_avg:159.78ms
step:83/1530 train_loss:4.7621 train_time:11664ms step_avg:159.78ms
step:84/1530 train_loss:4.7519 train_time:11824ms step_avg:159.79ms
step:85/1530 train_loss:4.6049 train_time:11984ms step_avg:159.78ms
step:86/1530 train_loss:4.8165 train_time:12144ms step_avg:159.79ms
step:87/1530 train_loss:4.7392 train_time:12304ms step_avg:159.79ms
step:88/1530 train_loss:4.7296 train_time:12464ms step_avg:159.80ms
step:89/1530 train_loss:4.6796 train_time:12626ms step_avg:159.82ms
step:90/1530 train_loss:4.6292 train_time:12785ms step_avg:159.81ms
step:91/1530 train_loss:4.6072 train_time:12946ms step_avg:159.83ms
step:92/1530 train_loss:4.7553 train_time:13106ms step_avg:159.83ms
step:93/1530 train_loss:4.5767 train_time:13267ms step_avg:159.84ms
step:94/1530 train_loss:4.6367 train_time:13428ms step_avg:159.86ms
step:95/1530 train_loss:4.6698 train_time:13588ms step_avg:159.86ms
step:96/1530 train_loss:4.5816 train_time:13749ms step_avg:159.88ms
step:97/1530 train_loss:4.6178 train_time:13910ms step_avg:159.88ms
step:98/1530 train_loss:4.5652 train_time:14071ms step_avg:159.90ms
step:99/1530 train_loss:4.6409 train_time:14232ms step_avg:159.91ms
step:100/1530 train_loss:4.6639 train_time:14392ms step_avg:159.91ms
step:101/1530 train_loss:4.5264 train_time:14553ms step_avg:159.92ms
step:102/1530 train_loss:4.6928 train_time:14714ms step_avg:159.93ms
step:103/1530 train_loss:4.5440 train_time:14875ms step_avg:159.94ms
step:104/1530 train_loss:4.5322 train_time:15035ms step_avg:159.95ms
step:105/1530 train_loss:4.5263 train_time:15195ms step_avg:159.95ms
step:106/1530 train_loss:4.5925 train_time:15355ms step_avg:159.94ms
step:107/1530 train_loss:4.4933 train_time:15514ms step_avg:159.94ms
step:108/1530 train_loss:4.3545 train_time:15675ms step_avg:159.95ms
step:109/1530 train_loss:4.4786 train_time:15836ms step_avg:159.96ms
step:110/1530 train_loss:4.4725 train_time:15996ms step_avg:159.96ms
step:111/1530 train_loss:4.4171 train_time:16156ms step_avg:159.96ms
step:112/1530 train_loss:4.5869 train_time:16317ms step_avg:159.97ms
step:113/1530 train_loss:4.4810 train_time:16477ms step_avg:159.97ms
step:114/1530 train_loss:4.3606 train_time:16638ms step_avg:159.98ms
step:115/1530 train_loss:4.4890 train_time:16802ms step_avg:160.02ms
step:116/1530 train_loss:4.4560 train_time:16966ms step_avg:160.06ms
step:117/1530 train_loss:4.3467 train_time:17129ms step_avg:160.09ms
step:118/1530 train_loss:4.5728 train_time:17293ms step_avg:160.12ms
step:119/1530 train_loss:4.4398 train_time:17457ms step_avg:160.16ms
step:120/1530 train_loss:4.3074 train_time:17623ms step_avg:160.21ms
step:121/1530 train_loss:4.2856 train_time:17786ms step_avg:160.23ms
step:122/1530 train_loss:4.4427 train_time:17949ms step_avg:160.26ms
step:123/1530 train_loss:4.2666 train_time:18113ms step_avg:160.29ms
step:124/1530 train_loss:4.5795 train_time:18278ms step_avg:160.33ms
step:125/1530 train_loss:4.4444 train_time:18442ms step_avg:160.37ms
step:125/1530 val_loss:4.3937 train_time:18488ms step_avg:160.77ms
step:126/1530 train_loss:4.4071 train_time:18608ms step_avg:160.42ms
step:127/1530 train_loss:4.4183 train_time:18774ms step_avg:160.46ms
step:128/1530 train_loss:4.3646 train_time:18939ms step_avg:160.50ms
step:129/1530 train_loss:4.6829 train_time:19104ms step_avg:160.54ms
step:130/1530 train_loss:4.3541 train_time:19267ms step_avg:160.56ms
step:131/1530 train_loss:4.3859 train_time:19431ms step_avg:160.58ms
step:132/1530 train_loss:4.3253 train_time:19595ms step_avg:160.61ms
step:133/1530 train_loss:4.4365 train_time:19759ms step_avg:160.64ms
step:134/1530 train_loss:4.2533 train_time:19924ms step_avg:160.68ms
step:135/1530 train_loss:4.4335 train_time:20088ms step_avg:160.70ms
step:136/1530 train_loss:4.2009 train_time:20251ms step_avg:160.72ms
step:137/1530 train_loss:4.3579 train_time:20416ms step_avg:160.75ms
step:138/1530 train_loss:4.2661 train_time:20580ms step_avg:160.78ms
step:139/1530 train_loss:4.3713 train_time:20744ms step_avg:160.81ms
step:140/1530 train_loss:4.4664 train_time:20908ms step_avg:160.83ms
step:141/1530 train_loss:4.3061 train_time:21071ms step_avg:160.85ms
step:142/1530 train_loss:4.3099 train_time:21236ms step_avg:160.88ms
step:143/1530 train_loss:4.2497 train_time:21401ms step_avg:160.91ms
step:144/1530 train_loss:4.3401 train_time:21565ms step_avg:160.93ms
step:145/1530 train_loss:4.2916 train_time:21729ms step_avg:160.95ms
step:146/1530 train_loss:4.1612 train_time:21892ms step_avg:160.97ms
step:147/1530 train_loss:4.3137 train_time:22056ms step_avg:160.99ms
step:148/1530 train_loss:4.3458 train_time:22221ms step_avg:161.02ms
step:149/1530 train_loss:4.3023 train_time:22386ms step_avg:161.05ms
step:150/1530 train_loss:4.4265 train_time:22550ms step_avg:161.07ms
step:151/1530 train_loss:4.2615 train_time:22715ms step_avg:161.10ms
step:152/1530 train_loss:4.2749 train_time:22879ms step_avg:161.12ms
step:153/1530 train_loss:4.3659 train_time:23043ms step_avg:161.14ms
step:154/1530 train_loss:4.3700 train_time:23206ms step_avg:161.16ms
step:155/1530 train_loss:4.2594 train_time:23369ms step_avg:161.17ms
step:156/1530 train_loss:4.3416 train_time:23533ms step_avg:161.19ms
step:157/1530 train_loss:4.3876 train_time:23698ms step_avg:161.21ms
step:158/1530 train_loss:4.2391 train_time:23862ms step_avg:161.23ms
step:159/1530 train_loss:4.3154 train_time:24026ms step_avg:161.25ms
step:160/1530 train_loss:4.1197 train_time:24190ms step_avg:161.26ms
step:161/1530 train_loss:4.3440 train_time:24353ms step_avg:161.28ms
step:162/1530 train_loss:4.3582 train_time:24517ms step_avg:161.29ms
step:163/1530 train_loss:4.3335 train_time:24680ms step_avg:161.31ms
step:164/1530 train_loss:4.1841 train_time:24844ms step_avg:161.32ms
step:165/1530 train_loss:4.2849 train_time:25008ms step_avg:161.34ms
step:166/1530 train_loss:4.3382 train_time:25171ms step_avg:161.35ms
step:167/1530 train_loss:4.2029 train_time:25335ms step_avg:161.37ms
step:168/1530 train_loss:4.2887 train_time:25500ms step_avg:161.39ms
step:169/1530 train_loss:4.1627 train_time:25664ms step_avg:161.41ms
step:170/1530 train_loss:4.0208 train_time:25829ms step_avg:161.43ms
step:171/1530 train_loss:4.1886 train_time:25992ms step_avg:161.44ms
step:172/1530 train_loss:4.2115 train_time:26155ms step_avg:161.45ms
step:173/1530 train_loss:4.2756 train_time:26319ms step_avg:161.46ms
step:174/1530 train_loss:4.4339 train_time:26482ms step_avg:161.47ms
step:175/1530 train_loss:4.2472 train_time:26645ms step_avg:161.48ms
step:176/1530 train_loss:4.0950 train_time:26808ms step_avg:161.49ms
step:177/1530 train_loss:4.0616 train_time:26969ms step_avg:161.49ms
step:178/1530 train_loss:4.1714 train_time:27133ms step_avg:161.50ms
step:179/1530 train_loss:4.1136 train_time:27297ms step_avg:161.52ms
step:180/1530 train_loss:4.1041 train_time:27459ms step_avg:161.52ms
step:181/1530 train_loss:4.2900 train_time:27623ms step_avg:161.54ms
step:182/1530 train_loss:4.1465 train_time:27785ms step_avg:161.54ms
step:183/1530 train_loss:4.1238 train_time:27947ms step_avg:161.55ms
step:184/1530 train_loss:4.1221 train_time:28111ms step_avg:161.56ms
step:185/1530 train_loss:4.2006 train_time:28274ms step_avg:161.57ms
step:186/1530 train_loss:4.1656 train_time:28437ms step_avg:161.57ms
step:187/1530 train_loss:4.2341 train_time:28600ms step_avg:161.58ms
step:188/1530 train_loss:4.1727 train_time:28897ms step_avg:162.35ms
step:189/1530 train_loss:4.1120 train_time:29230ms step_avg:163.30ms
step:190/1530 train_loss:4.2119 train_time:29392ms step_avg:163.29ms
step:191/1530 train_loss:4.0788 train_time:29553ms step_avg:163.28ms
step:192/1530 train_loss:4.0312 train_time:29718ms step_avg:163.29ms
step:193/1530 train_loss:4.2559 train_time:29882ms step_avg:163.29ms
step:194/1530 train_loss:4.1701 train_time:30044ms step_avg:163.29ms
step:195/1530 train_loss:4.3565 train_time:30208ms step_avg:163.28ms
step:196/1530 train_loss:4.1807 train_time:30370ms step_avg:163.28ms
step:197/1530 train_loss:4.0407 train_time:30533ms step_avg:163.28ms
step:198/1530 train_loss:4.1757 train_time:30698ms step_avg:163.29ms
step:199/1530 train_loss:4.0383 train_time:30861ms step_avg:163.28ms
step:200/1530 train_loss:4.1076 train_time:31024ms step_avg:163.29ms
step:201/1530 train_loss:3.9940 train_time:31186ms step_avg:163.28ms
step:202/1530 train_loss:4.2426 train_time:31349ms step_avg:163.27ms
step:203/1530 train_loss:4.0614 train_time:31511ms step_avg:163.27ms
step:204/1530 train_loss:4.1827 train_time:31675ms step_avg:163.27ms
step:205/1530 train_loss:4.2401 train_time:31839ms step_avg:163.28ms
step:206/1530 train_loss:3.9473 train_time:32002ms step_avg:163.28ms
step:207/1530 train_loss:4.0813 train_time:32165ms step_avg:163.27ms
step:208/1530 train_loss:4.0948 train_time:32328ms step_avg:163.27ms
step:209/1530 train_loss:4.2357 train_time:32490ms step_avg:163.27ms
step:210/1530 train_loss:4.1754 train_time:32654ms step_avg:163.27ms
step:211/1530 train_loss:4.0578 train_time:32817ms step_avg:163.27ms
step:212/1530 train_loss:4.1118 train_time:32981ms step_avg:163.27ms
step:213/1530 train_loss:4.0528 train_time:33143ms step_avg:163.27ms
step:214/1530 train_loss:4.1181 train_time:33307ms step_avg:163.27ms
step:215/1530 train_loss:3.9619 train_time:33469ms step_avg:163.26ms
step:216/1530 train_loss:4.0044 train_time:33632ms step_avg:163.26ms
step:217/1530 train_loss:4.0215 train_time:33796ms step_avg:163.26ms
step:218/1530 train_loss:4.0830 train_time:33958ms step_avg:163.26ms
step:219/1530 train_loss:4.0748 train_time:34123ms step_avg:163.27ms
step:220/1530 train_loss:4.0815 train_time:34286ms step_avg:163.26ms
step:221/1530 train_loss:4.0859 train_time:34448ms step_avg:163.26ms
step:222/1530 train_loss:3.9963 train_time:34612ms step_avg:163.26ms
step:223/1530 train_loss:3.9865 train_time:34775ms step_avg:163.26ms
step:224/1530 train_loss:4.2910 train_time:34937ms step_avg:163.26ms
step:225/1530 train_loss:3.9183 train_time:35101ms step_avg:163.26ms
step:226/1530 train_loss:3.9880 train_time:35264ms step_avg:163.26ms
step:227/1530 train_loss:3.9710 train_time:35427ms step_avg:163.26ms
step:228/1530 train_loss:4.1510 train_time:35591ms step_avg:163.26ms
step:229/1530 train_loss:3.9352 train_time:35757ms step_avg:163.28ms
step:230/1530 train_loss:4.0309 train_time:35925ms step_avg:163.29ms
step:231/1530 train_loss:3.9017 train_time:36091ms step_avg:163.31ms
step:232/1530 train_loss:3.9660 train_time:36258ms step_avg:163.32ms
step:233/1530 train_loss:4.0832 train_time:36425ms step_avg:163.34ms
step:234/1530 train_loss:4.0267 train_time:36591ms step_avg:163.35ms
step:235/1530 train_loss:3.9028 train_time:36759ms step_avg:163.37ms
step:236/1530 train_loss:4.0842 train_time:36926ms step_avg:163.39ms
step:237/1530 train_loss:4.0839 train_time:37091ms step_avg:163.40ms
step:238/1530 train_loss:3.9413 train_time:37258ms step_avg:163.41ms
step:239/1530 train_loss:4.0780 train_time:37425ms step_avg:163.43ms
step:240/1530 train_loss:4.1136 train_time:37591ms step_avg:163.44ms
step:241/1530 train_loss:3.9656 train_time:37758ms step_avg:163.45ms
step:242/1530 train_loss:4.1461 train_time:37925ms step_avg:163.47ms
step:243/1530 train_loss:4.0128 train_time:38091ms step_avg:163.48ms
step:244/1530 train_loss:4.0827 train_time:38257ms step_avg:163.49ms
step:245/1530 train_loss:4.1406 train_time:38423ms step_avg:163.50ms
step:246/1530 train_loss:4.0583 train_time:38589ms step_avg:163.51ms
step:247/1530 train_loss:4.0064 train_time:38754ms step_avg:163.52ms
step:248/1530 train_loss:4.1036 train_time:38922ms step_avg:163.54ms
step:249/1530 train_loss:3.9220 train_time:39087ms step_avg:163.54ms
step:250/1530 train_loss:3.9725 train_time:39253ms step_avg:163.55ms
step:250/1530 val_loss:4.0045 train_time:39302ms step_avg:163.76ms
step:251/1530 train_loss:4.0787 train_time:39423ms step_avg:163.58ms
step:252/1530 train_loss:4.1886 train_time:39592ms step_avg:163.60ms
step:253/1530 train_loss:3.9362 train_time:39758ms step_avg:163.61ms
step:254/1530 train_loss:3.8792 train_time:39924ms step_avg:163.62ms
step:255/1530 train_loss:4.0767 train_time:40090ms step_avg:163.63ms
step:256/1530 train_loss:3.9888 train_time:40254ms step_avg:163.64ms
step:257/1530 train_loss:3.9924 train_time:40420ms step_avg:163.65ms
step:258/1530 train_loss:3.9815 train_time:40588ms step_avg:163.66ms
step:259/1530 train_loss:4.0262 train_time:40753ms step_avg:163.67ms
step:260/1530 train_loss:4.0512 train_time:40919ms step_avg:163.68ms
step:261/1530 train_loss:4.0245 train_time:41089ms step_avg:163.70ms
step:262/1530 train_loss:3.9946 train_time:41254ms step_avg:163.71ms
step:263/1530 train_loss:3.8928 train_time:41421ms step_avg:163.72ms
step:264/1530 train_loss:3.9874 train_time:41588ms step_avg:163.73ms
step:265/1530 train_loss:3.8753 train_time:41754ms step_avg:163.74ms
step:266/1530 train_loss:3.9242 train_time:41920ms step_avg:163.75ms
step:267/1530 train_loss:3.9368 train_time:42088ms step_avg:163.77ms
step:268/1530 train_loss:3.9642 train_time:42254ms step_avg:163.77ms
step:269/1530 train_loss:3.8591 train_time:42419ms step_avg:163.78ms
step:270/1530 train_loss:4.0960 train_time:42586ms step_avg:163.79ms
step:271/1530 train_loss:3.9701 train_time:42752ms step_avg:163.80ms
step:272/1530 train_loss:3.9242 train_time:42918ms step_avg:163.81ms
step:273/1530 train_loss:3.9399 train_time:43086ms step_avg:163.82ms
step:274/1530 train_loss:4.0330 train_time:43252ms step_avg:163.83ms
step:275/1530 train_loss:4.0602 train_time:43418ms step_avg:163.84ms
step:276/1530 train_loss:4.2247 train_time:43587ms step_avg:163.86ms
step:277/1530 train_loss:4.0347 train_time:43752ms step_avg:163.87ms
step:278/1530 train_loss:4.0849 train_time:43919ms step_avg:163.88ms
step:279/1530 train_loss:4.0039 train_time:44087ms step_avg:163.89ms
step:280/1530 train_loss:4.1819 train_time:44253ms step_avg:163.90ms
step:281/1530 train_loss:3.9719 train_time:44419ms step_avg:163.91ms
step:282/1530 train_loss:3.9401 train_time:44589ms step_avg:163.93ms
step:283/1530 train_loss:3.9124 train_time:44756ms step_avg:163.94ms
step:284/1530 train_loss:4.0448 train_time:44922ms step_avg:163.95ms
step:285/1530 train_loss:4.0598 train_time:45089ms step_avg:163.96ms
step:286/1530 train_loss:4.0902 train_time:45253ms step_avg:163.96ms
step:287/1530 train_loss:3.9030 train_time:45419ms step_avg:163.97ms
step:288/1530 train_loss:4.0183 train_time:45586ms step_avg:163.98ms
step:289/1530 train_loss:3.8735 train_time:45751ms step_avg:163.98ms
step:290/1530 train_loss:3.8605 train_time:45916ms step_avg:163.99ms
step:291/1530 train_loss:3.9066 train_time:46082ms step_avg:163.99ms
step:292/1530 train_loss:3.8673 train_time:46246ms step_avg:163.99ms
step:293/1530 train_loss:3.9015 train_time:46411ms step_avg:164.00ms
step:294/1530 train_loss:3.9320 train_time:46576ms step_avg:164.00ms
step:295/1530 train_loss:3.8398 train_time:46741ms step_avg:164.00ms
step:296/1530 train_loss:3.8609 train_time:46907ms step_avg:164.01ms
step:297/1530 train_loss:3.8730 train_time:47073ms step_avg:164.02ms
step:298/1530 train_loss:3.9728 train_time:47237ms step_avg:164.02ms
step:299/1530 train_loss:3.8202 train_time:47403ms step_avg:164.02ms
step:300/1530 train_loss:3.9686 train_time:47568ms step_avg:164.03ms
step:301/1530 train_loss:3.9657 train_time:47732ms step_avg:164.03ms
step:302/1530 train_loss:3.9319 train_time:47898ms step_avg:164.03ms
step:303/1530 train_loss:3.9729 train_time:48065ms step_avg:164.04ms
step:304/1530 train_loss:3.9669 train_time:48229ms step_avg:164.05ms
step:305/1530 train_loss:4.4489 train_time:48394ms step_avg:164.05ms
step:306/1530 train_loss:3.9328 train_time:48559ms step_avg:164.05ms
step:307/1530 train_loss:3.8366 train_time:48725ms step_avg:164.06ms
step:308/1530 train_loss:3.9734 train_time:48890ms step_avg:164.06ms
step:309/1530 train_loss:3.8731 train_time:49055ms step_avg:164.06ms
step:310/1530 train_loss:4.0853 train_time:49220ms step_avg:164.07ms
step:311/1530 train_loss:3.9268 train_time:49387ms step_avg:164.08ms
step:312/1530 train_loss:3.8697 train_time:49552ms step_avg:164.08ms
step:313/1530 train_loss:3.9362 train_time:49717ms step_avg:164.08ms
step:314/1530 train_loss:4.0613 train_time:49883ms step_avg:164.09ms
step:315/1530 train_loss:3.9446 train_time:50048ms step_avg:164.09ms
step:316/1530 train_loss:3.8004 train_time:50213ms step_avg:164.09ms
step:317/1530 train_loss:3.8774 train_time:50378ms step_avg:164.10ms
step:318/1530 train_loss:3.9260 train_time:50543ms step_avg:164.10ms
step:319/1530 train_loss:3.8858 train_time:50708ms step_avg:164.10ms
step:320/1530 train_loss:4.0097 train_time:50873ms step_avg:164.11ms
step:321/1530 train_loss:3.9568 train_time:51039ms step_avg:164.11ms
step:322/1530 train_loss:3.9364 train_time:51205ms step_avg:164.12ms
step:323/1530 train_loss:4.0140 train_time:51371ms step_avg:164.12ms
step:324/1530 train_loss:3.9482 train_time:51535ms step_avg:164.12ms
step:325/1530 train_loss:4.0160 train_time:51700ms step_avg:164.13ms
step:326/1530 train_loss:3.8956 train_time:51868ms step_avg:164.14ms
step:327/1530 train_loss:4.3992 train_time:52033ms step_avg:164.14ms
step:328/1530 train_loss:4.0804 train_time:52199ms step_avg:164.15ms
step:329/1530 train_loss:3.7943 train_time:52367ms step_avg:164.16ms
step:330/1530 train_loss:3.7485 train_time:52531ms step_avg:164.16ms
step:331/1530 train_loss:3.9750 train_time:52696ms step_avg:164.16ms
step:332/1530 train_loss:3.9175 train_time:52861ms step_avg:164.17ms
step:333/1530 train_loss:3.8851 train_time:53027ms step_avg:164.17ms
step:334/1530 train_loss:3.8413 train_time:53193ms step_avg:164.17ms
step:335/1530 train_loss:4.0190 train_time:53357ms step_avg:164.18ms
step:336/1530 train_loss:3.9568 train_time:53522ms step_avg:164.18ms
step:337/1530 train_loss:4.4309 train_time:53690ms step_avg:164.19ms
step:338/1530 train_loss:3.9375 train_time:53855ms step_avg:164.19ms
step:339/1530 train_loss:3.8632 train_time:54021ms step_avg:164.20ms
step:340/1530 train_loss:3.9409 train_time:54188ms step_avg:164.21ms
step:341/1530 train_loss:3.8529 train_time:54355ms step_avg:164.21ms
step:342/1530 train_loss:3.8112 train_time:54522ms step_avg:164.22ms
step:343/1530 train_loss:3.8382 train_time:54692ms step_avg:164.24ms
step:344/1530 train_loss:3.9968 train_time:54859ms step_avg:164.25ms
step:345/1530 train_loss:3.8188 train_time:55029ms step_avg:164.27ms
step:346/1530 train_loss:3.7653 train_time:55197ms step_avg:164.28ms
step:347/1530 train_loss:3.8010 train_time:55366ms step_avg:164.29ms
step:348/1530 train_loss:3.8600 train_time:55533ms step_avg:164.30ms
step:349/1530 train_loss:3.8265 train_time:55702ms step_avg:164.31ms
step:350/1530 train_loss:3.5748 train_time:55872ms step_avg:164.33ms
step:351/1530 train_loss:3.8274 train_time:56039ms step_avg:164.34ms
step:352/1530 train_loss:4.1844 train_time:56208ms step_avg:164.35ms
step:353/1530 train_loss:3.6567 train_time:56376ms step_avg:164.36ms
step:354/1530 train_loss:3.9234 train_time:56544ms step_avg:164.37ms
step:355/1530 train_loss:3.7827 train_time:56713ms step_avg:164.38ms
step:356/1530 train_loss:3.8858 train_time:56878ms step_avg:164.39ms
step:357/1530 train_loss:3.7630 train_time:57049ms step_avg:164.41ms
step:358/1530 train_loss:3.8606 train_time:57218ms step_avg:164.42ms
step:359/1530 train_loss:3.7748 train_time:57389ms step_avg:164.44ms
step:360/1530 train_loss:3.4332 train_time:57558ms step_avg:164.45ms
step:361/1530 train_loss:4.0286 train_time:57726ms step_avg:164.46ms
step:362/1530 train_loss:3.9170 train_time:57894ms step_avg:164.47ms
step:363/1530 train_loss:3.8414 train_time:58062ms step_avg:164.48ms
step:364/1530 train_loss:3.7544 train_time:58230ms step_avg:164.49ms
step:365/1530 train_loss:3.9149 train_time:58397ms step_avg:164.50ms
step:366/1530 train_loss:3.8713 train_time:58566ms step_avg:164.51ms
step:367/1530 train_loss:3.8573 train_time:58734ms step_avg:164.52ms
step:368/1530 train_loss:3.8502 train_time:58901ms step_avg:164.53ms
step:369/1530 train_loss:3.7469 train_time:59071ms step_avg:164.54ms
step:370/1530 train_loss:3.8768 train_time:59238ms step_avg:164.55ms
step:371/1530 train_loss:3.7307 train_time:59408ms step_avg:164.56ms
step:372/1530 train_loss:3.6971 train_time:59577ms step_avg:164.58ms
step:373/1530 train_loss:3.9092 train_time:59744ms step_avg:164.58ms
step:374/1530 train_loss:3.8318 train_time:59912ms step_avg:164.59ms
step:375/1530 train_loss:3.8046 train_time:60080ms step_avg:164.60ms
step:375/1530 val_loss:3.8271 train_time:60129ms step_avg:164.74ms
step:376/1530 train_loss:3.8665 train_time:60251ms step_avg:164.62ms
step:377/1530 train_loss:3.7869 train_time:60556ms step_avg:165.00ms
step:378/1530 train_loss:3.8497 train_time:60734ms step_avg:165.04ms
step:379/1530 train_loss:3.8692 train_time:61053ms step_avg:165.46ms
step:380/1530 train_loss:3.9540 train_time:61221ms step_avg:165.46ms
step:381/1530 train_loss:3.8491 train_time:61388ms step_avg:165.47ms
step:382/1530 train_loss:3.8117 train_time:61558ms step_avg:165.48ms
step:383/1530 train_loss:3.8017 train_time:61726ms step_avg:165.49ms
step:384/1530 train_loss:3.8754 train_time:61893ms step_avg:165.49ms
step:385/1530 train_loss:3.7933 train_time:62063ms step_avg:165.50ms
step:386/1530 train_loss:3.8970 train_time:62230ms step_avg:165.51ms
step:387/1530 train_loss:4.0648 train_time:62399ms step_avg:165.51ms
step:388/1530 train_loss:3.7985 train_time:62566ms step_avg:165.52ms
step:389/1530 train_loss:3.7977 train_time:62735ms step_avg:165.53ms
step:390/1530 train_loss:3.9053 train_time:62904ms step_avg:165.54ms
step:391/1530 train_loss:3.8180 train_time:63071ms step_avg:165.54ms
step:392/1530 train_loss:3.9273 train_time:63239ms step_avg:165.55ms
step:393/1530 train_loss:3.7667 train_time:63406ms step_avg:165.55ms
step:394/1530 train_loss:3.8877 train_time:63574ms step_avg:165.56ms
step:395/1530 train_loss:3.6348 train_time:63743ms step_avg:165.57ms
step:396/1530 train_loss:3.8453 train_time:63910ms step_avg:165.57ms
step:397/1530 train_loss:3.8603 train_time:64079ms step_avg:165.58ms
step:398/1530 train_loss:3.8906 train_time:64246ms step_avg:165.58ms
step:399/1530 train_loss:3.7718 train_time:64413ms step_avg:165.59ms
step:400/1530 train_loss:3.8312 train_time:64581ms step_avg:165.59ms
step:401/1530 train_loss:3.9075 train_time:64748ms step_avg:165.60ms
step:402/1530 train_loss:3.8496 train_time:64917ms step_avg:165.60ms
step:403/1530 train_loss:3.9627 train_time:65083ms step_avg:165.61ms
step:404/1530 train_loss:3.6804 train_time:65250ms step_avg:165.61ms
step:405/1530 train_loss:3.7824 train_time:65419ms step_avg:165.62ms
step:406/1530 train_loss:4.0947 train_time:65585ms step_avg:165.62ms
step:407/1530 train_loss:3.7812 train_time:65753ms step_avg:165.62ms
step:408/1530 train_loss:3.8200 train_time:65920ms step_avg:165.63ms
step:409/1530 train_loss:3.8577 train_time:66086ms step_avg:165.63ms
step:410/1530 train_loss:3.7611 train_time:66255ms step_avg:165.64ms
step:411/1530 train_loss:3.7649 train_time:66422ms step_avg:165.64ms
step:412/1530 train_loss:4.1704 train_time:66589ms step_avg:165.64ms
step:413/1530 train_loss:3.6309 train_time:66756ms step_avg:165.65ms
step:414/1530 train_loss:4.0058 train_time:66923ms step_avg:165.65ms
step:415/1530 train_loss:3.7447 train_time:67090ms step_avg:165.65ms
step:416/1530 train_loss:3.7618 train_time:67258ms step_avg:165.66ms
step:417/1530 train_loss:3.9522 train_time:67426ms step_avg:165.67ms
step:418/1530 train_loss:3.6967 train_time:67592ms step_avg:165.67ms
step:419/1530 train_loss:3.8069 train_time:67761ms step_avg:165.67ms
step:420/1530 train_loss:3.7064 train_time:67928ms step_avg:165.68ms
step:421/1530 train_loss:3.6521 train_time:68095ms step_avg:165.68ms
step:422/1530 train_loss:3.7835 train_time:68262ms step_avg:165.68ms
step:423/1530 train_loss:3.8754 train_time:68428ms step_avg:165.69ms
step:424/1530 train_loss:3.6196 train_time:68597ms step_avg:165.69ms
step:425/1530 train_loss:3.7918 train_time:68764ms step_avg:165.70ms
step:426/1530 train_loss:3.6520 train_time:68931ms step_avg:165.70ms
step:427/1530 train_loss:3.8937 train_time:69100ms step_avg:165.71ms
step:428/1530 train_loss:3.8175 train_time:69267ms step_avg:165.71ms
step:429/1530 train_loss:3.7627 train_time:69435ms step_avg:165.72ms
step:430/1530 train_loss:3.7044 train_time:69602ms step_avg:165.72ms
step:431/1530 train_loss:3.6362 train_time:69769ms step_avg:165.72ms
step:432/1530 train_loss:3.7649 train_time:69938ms step_avg:165.73ms
step:433/1530 train_loss:3.8203 train_time:70105ms step_avg:165.73ms
step:434/1530 train_loss:3.7746 train_time:70270ms step_avg:165.73ms
step:435/1530 train_loss:3.8084 train_time:70440ms step_avg:165.74ms
step:436/1530 train_loss:3.8352 train_time:70607ms step_avg:165.74ms
step:437/1530 train_loss:3.7271 train_time:70774ms step_avg:165.75ms
step:438/1530 train_loss:3.7074 train_time:70942ms step_avg:165.75ms
step:439/1530 train_loss:3.7087 train_time:71108ms step_avg:165.75ms
step:440/1530 train_loss:3.8896 train_time:71276ms step_avg:165.76ms
step:441/1530 train_loss:3.7597 train_time:71445ms step_avg:165.77ms
step:442/1530 train_loss:3.7317 train_time:71612ms step_avg:165.77ms
step:443/1530 train_loss:3.6229 train_time:71780ms step_avg:165.77ms
step:444/1530 train_loss:3.9243 train_time:71947ms step_avg:165.78ms
step:445/1530 train_loss:3.8464 train_time:72113ms step_avg:165.78ms
step:446/1530 train_loss:3.8381 train_time:72281ms step_avg:165.78ms
step:447/1530 train_loss:3.7504 train_time:72448ms step_avg:165.78ms
step:448/1530 train_loss:3.8554 train_time:72618ms step_avg:165.79ms
step:449/1530 train_loss:3.6871 train_time:72784ms step_avg:165.79ms
step:450/1530 train_loss:3.7196 train_time:72951ms step_avg:165.80ms
step:451/1530 train_loss:3.5814 train_time:73119ms step_avg:165.80ms
step:452/1530 train_loss:3.7130 train_time:73285ms step_avg:165.80ms
step:453/1530 train_loss:3.6762 train_time:73452ms step_avg:165.81ms
step:454/1530 train_loss:3.6442 train_time:73621ms step_avg:165.81ms
step:455/1530 train_loss:3.8439 train_time:73788ms step_avg:165.82ms
step:456/1530 train_loss:3.7233 train_time:73958ms step_avg:165.83ms
step:457/1530 train_loss:3.7854 train_time:74128ms step_avg:165.83ms
step:458/1530 train_loss:3.8305 train_time:74297ms step_avg:165.84ms
step:459/1530 train_loss:3.6342 train_time:74467ms step_avg:165.85ms
step:460/1530 train_loss:3.7888 train_time:74637ms step_avg:165.86ms
step:461/1530 train_loss:3.6984 train_time:74807ms step_avg:165.87ms
step:462/1530 train_loss:3.7392 train_time:74977ms step_avg:165.88ms
step:463/1530 train_loss:3.7795 train_time:75147ms step_avg:165.89ms
step:464/1530 train_loss:3.7190 train_time:75316ms step_avg:165.90ms
step:465/1530 train_loss:3.7166 train_time:75484ms step_avg:165.90ms
step:466/1530 train_loss:3.7928 train_time:75653ms step_avg:165.91ms
step:467/1530 train_loss:3.8251 train_time:75826ms step_avg:165.92ms
step:468/1530 train_loss:3.7906 train_time:75995ms step_avg:165.93ms
step:469/1530 train_loss:3.6857 train_time:76165ms step_avg:165.94ms
step:470/1530 train_loss:3.7661 train_time:76333ms step_avg:165.94ms
step:471/1530 train_loss:3.8098 train_time:76503ms step_avg:165.95ms
step:472/1530 train_loss:3.7851 train_time:76674ms step_avg:165.96ms
step:473/1530 train_loss:3.7152 train_time:76845ms step_avg:165.97ms
step:474/1530 train_loss:3.5937 train_time:77014ms step_avg:165.98ms
step:475/1530 train_loss:4.0137 train_time:77183ms step_avg:165.98ms
step:476/1530 train_loss:3.7588 train_time:77352ms step_avg:165.99ms
step:477/1530 train_loss:3.5949 train_time:77524ms step_avg:166.00ms
step:478/1530 train_loss:3.8256 train_time:77694ms step_avg:166.01ms
step:479/1530 train_loss:3.7802 train_time:77864ms step_avg:166.02ms
step:480/1530 train_loss:3.9168 train_time:78032ms step_avg:166.03ms
step:481/1530 train_loss:3.7305 train_time:78204ms step_avg:166.04ms
step:482/1530 train_loss:3.5310 train_time:78373ms step_avg:166.05ms
step:483/1530 train_loss:3.8075 train_time:78543ms step_avg:166.05ms
step:484/1530 train_loss:3.6607 train_time:78715ms step_avg:166.06ms
step:485/1530 train_loss:3.6567 train_time:78884ms step_avg:166.07ms
step:486/1530 train_loss:3.5741 train_time:79053ms step_avg:166.08ms
step:487/1530 train_loss:3.6853 train_time:79224ms step_avg:166.09ms
step:488/1530 train_loss:3.8755 train_time:79393ms step_avg:166.09ms
step:489/1530 train_loss:3.7143 train_time:79564ms step_avg:166.10ms
step:490/1530 train_loss:3.5926 train_time:79733ms step_avg:166.11ms
step:491/1530 train_loss:3.6186 train_time:79902ms step_avg:166.12ms
step:492/1530 train_loss:3.7345 train_time:80072ms step_avg:166.12ms
step:493/1530 train_loss:3.5807 train_time:80243ms step_avg:166.14ms
step:494/1530 train_loss:3.7142 train_time:80411ms step_avg:166.14ms
step:495/1530 train_loss:3.6651 train_time:80583ms step_avg:166.15ms
step:496/1530 train_loss:3.5141 train_time:80753ms step_avg:166.16ms
step:497/1530 train_loss:3.7361 train_time:80923ms step_avg:166.17ms
step:498/1530 train_loss:3.7908 train_time:81091ms step_avg:166.17ms
step:499/1530 train_loss:3.8192 train_time:81263ms step_avg:166.18ms
step:500/1530 train_loss:3.7376 train_time:81433ms step_avg:166.19ms
step:500/1530 val_loss:3.7066 train_time:81481ms step_avg:166.29ms
step:501/1530 train_loss:3.8019 train_time:81604ms step_avg:166.20ms
step:502/1530 train_loss:3.7498 train_time:81775ms step_avg:166.21ms
step:503/1530 train_loss:3.7742 train_time:81945ms step_avg:166.22ms
step:504/1530 train_loss:3.7207 train_time:82113ms step_avg:166.22ms
step:505/1530 train_loss:3.8110 train_time:82284ms step_avg:166.23ms
step:506/1530 train_loss:3.6501 train_time:82453ms step_avg:166.24ms
step:507/1530 train_loss:3.7665 train_time:82622ms step_avg:166.24ms
step:508/1530 train_loss:3.8251 train_time:82792ms step_avg:166.25ms
step:509/1530 train_loss:3.7770 train_time:82962ms step_avg:166.26ms
step:510/1530 train_loss:3.5851 train_time:83133ms step_avg:166.27ms
step:511/1530 train_loss:3.7793 train_time:83304ms step_avg:166.28ms
step:512/1530 train_loss:3.7161 train_time:83477ms step_avg:166.29ms
step:513/1530 train_loss:3.6649 train_time:83644ms step_avg:166.29ms
step:514/1530 train_loss:3.7881 train_time:83814ms step_avg:166.30ms
step:515/1530 train_loss:3.7362 train_time:83983ms step_avg:166.30ms
step:516/1530 train_loss:4.0727 train_time:84153ms step_avg:166.31ms
step:517/1530 train_loss:3.6862 train_time:84322ms step_avg:166.32ms
step:518/1530 train_loss:3.7694 train_time:84490ms step_avg:166.32ms
step:519/1530 train_loss:3.6556 train_time:84660ms step_avg:166.33ms
step:520/1530 train_loss:3.6898 train_time:84828ms step_avg:166.33ms
step:521/1530 train_loss:3.6606 train_time:84998ms step_avg:166.34ms
step:522/1530 train_loss:3.6544 train_time:85168ms step_avg:166.34ms
step:523/1530 train_loss:4.3000 train_time:85338ms step_avg:166.35ms
step:524/1530 train_loss:3.7401 train_time:85507ms step_avg:166.36ms
step:525/1530 train_loss:3.6817 train_time:85677ms step_avg:166.36ms
step:526/1530 train_loss:3.7043 train_time:85845ms step_avg:166.37ms
step:527/1530 train_loss:3.6605 train_time:86012ms step_avg:166.37ms
step:528/1530 train_loss:3.6235 train_time:86183ms step_avg:166.38ms
step:529/1530 train_loss:3.8460 train_time:86352ms step_avg:166.38ms
step:530/1530 train_loss:3.6512 train_time:86521ms step_avg:166.39ms
step:531/1530 train_loss:3.9185 train_time:86689ms step_avg:166.39ms
step:532/1530 train_loss:3.7323 train_time:86861ms step_avg:166.40ms
step:533/1530 train_loss:3.6486 train_time:87028ms step_avg:166.40ms
step:534/1530 train_loss:3.6729 train_time:87198ms step_avg:166.41ms
step:535/1530 train_loss:3.6117 train_time:87367ms step_avg:166.41ms
step:536/1530 train_loss:3.7490 train_time:87537ms step_avg:166.42ms
step:537/1530 train_loss:3.7295 train_time:87708ms step_avg:166.43ms
step:538/1530 train_loss:3.6318 train_time:87881ms step_avg:166.44ms
step:539/1530 train_loss:4.1086 train_time:88051ms step_avg:166.45ms
step:540/1530 train_loss:3.6749 train_time:88220ms step_avg:166.45ms
step:541/1530 train_loss:3.7861 train_time:88388ms step_avg:166.45ms
step:542/1530 train_loss:3.5883 train_time:88558ms step_avg:166.46ms
step:543/1530 train_loss:3.5940 train_time:88726ms step_avg:166.46ms
step:544/1530 train_loss:3.6396 train_time:88894ms step_avg:166.47ms
step:545/1530 train_loss:3.5949 train_time:89064ms step_avg:166.48ms
step:546/1530 train_loss:3.6257 train_time:89232ms step_avg:166.48ms
step:547/1530 train_loss:3.6439 train_time:89402ms step_avg:166.48ms
step:548/1530 train_loss:3.6084 train_time:89570ms step_avg:166.49ms
step:549/1530 train_loss:3.7235 train_time:89739ms step_avg:166.49ms
step:550/1530 train_loss:3.6175 train_time:89908ms step_avg:166.50ms
step:551/1530 train_loss:3.6361 train_time:90079ms step_avg:166.50ms
step:552/1530 train_loss:3.9387 train_time:90248ms step_avg:166.51ms
step:553/1530 train_loss:3.7556 train_time:90417ms step_avg:166.51ms
step:554/1530 train_loss:3.7092 train_time:90586ms step_avg:166.52ms
step:555/1530 train_loss:3.6364 train_time:90755ms step_avg:166.52ms
step:556/1530 train_loss:3.7006 train_time:90923ms step_avg:166.53ms
step:557/1530 train_loss:3.3189 train_time:91092ms step_avg:166.53ms
step:558/1530 train_loss:3.6182 train_time:91263ms step_avg:166.54ms
step:559/1530 train_loss:3.6428 train_time:91431ms step_avg:166.54ms
step:560/1530 train_loss:3.6903 train_time:91602ms step_avg:166.55ms
step:561/1530 train_loss:3.6101 train_time:91770ms step_avg:166.55ms
step:562/1530 train_loss:3.5588 train_time:91940ms step_avg:166.56ms
step:563/1530 train_loss:3.7575 train_time:92108ms step_avg:166.56ms
step:564/1530 train_loss:3.5727 train_time:92280ms step_avg:166.57ms
step:565/1530 train_loss:3.6806 train_time:92449ms step_avg:166.57ms
step:566/1530 train_loss:3.6107 train_time:92751ms step_avg:166.82ms
step:567/1530 train_loss:3.5977 train_time:92933ms step_avg:166.85ms
step:568/1530 train_loss:3.6928 train_time:93103ms step_avg:166.85ms
step:569/1530 train_loss:3.6501 train_time:93428ms step_avg:167.13ms
step:570/1530 train_loss:3.6905 train_time:93598ms step_avg:167.14ms
step:571/1530 train_loss:3.7584 train_time:93768ms step_avg:167.14ms
step:572/1530 train_loss:3.7251 train_time:93940ms step_avg:167.15ms
step:573/1530 train_loss:3.7384 train_time:94113ms step_avg:167.16ms
step:574/1530 train_loss:3.7797 train_time:94287ms step_avg:167.18ms
step:575/1530 train_loss:3.7302 train_time:94459ms step_avg:167.18ms
step:576/1530 train_loss:3.7609 train_time:94629ms step_avg:167.19ms
step:577/1530 train_loss:3.6734 train_time:94802ms step_avg:167.20ms
step:578/1530 train_loss:3.6690 train_time:94973ms step_avg:167.21ms
step:579/1530 train_loss:3.6735 train_time:95145ms step_avg:167.21ms
step:580/1530 train_loss:3.5914 train_time:95315ms step_avg:167.22ms
step:581/1530 train_loss:3.6393 train_time:95487ms step_avg:167.23ms
step:582/1530 train_loss:3.8501 train_time:95658ms step_avg:167.23ms
step:583/1530 train_loss:3.6275 train_time:95829ms step_avg:167.24ms
step:584/1530 train_loss:3.5924 train_time:96002ms step_avg:167.25ms
step:585/1530 train_loss:3.7892 train_time:96173ms step_avg:167.26ms
step:586/1530 train_loss:3.5184 train_time:96345ms step_avg:167.27ms
step:587/1530 train_loss:3.6711 train_time:96515ms step_avg:167.27ms
step:588/1530 train_loss:3.6399 train_time:96686ms step_avg:167.28ms
step:589/1530 train_loss:3.9992 train_time:96856ms step_avg:167.28ms
step:590/1530 train_loss:3.7749 train_time:97027ms step_avg:167.29ms
step:591/1530 train_loss:3.5050 train_time:97199ms step_avg:167.30ms
step:592/1530 train_loss:3.5366 train_time:97375ms step_avg:167.31ms
step:593/1530 train_loss:3.5038 train_time:97547ms step_avg:167.32ms
step:594/1530 train_loss:3.5544 train_time:97719ms step_avg:167.33ms
step:595/1530 train_loss:3.9210 train_time:97892ms step_avg:167.34ms
step:596/1530 train_loss:3.6541 train_time:98065ms step_avg:167.35ms
step:597/1530 train_loss:3.5873 train_time:98236ms step_avg:167.35ms
step:598/1530 train_loss:3.6554 train_time:98407ms step_avg:167.36ms
step:599/1530 train_loss:3.4761 train_time:98579ms step_avg:167.37ms
step:600/1530 train_loss:3.5978 train_time:98750ms step_avg:167.37ms
step:601/1530 train_loss:3.6444 train_time:98924ms step_avg:167.38ms
step:602/1530 train_loss:3.6715 train_time:99097ms step_avg:167.39ms
step:603/1530 train_loss:3.7770 train_time:99268ms step_avg:167.40ms
step:604/1530 train_loss:3.6106 train_time:99440ms step_avg:167.41ms
step:605/1530 train_loss:3.6059 train_time:99612ms step_avg:167.42ms
step:606/1530 train_loss:3.5791 train_time:99786ms step_avg:167.43ms
step:607/1530 train_loss:3.8438 train_time:99958ms step_avg:167.43ms
step:608/1530 train_loss:3.6328 train_time:100129ms step_avg:167.44ms
step:609/1530 train_loss:3.6179 train_time:100300ms step_avg:167.45ms
step:610/1530 train_loss:3.7060 train_time:100470ms step_avg:167.45ms
step:611/1530 train_loss:3.5984 train_time:100641ms step_avg:167.46ms
step:612/1530 train_loss:3.5752 train_time:100812ms step_avg:167.46ms
step:613/1530 train_loss:3.7696 train_time:100984ms step_avg:167.47ms
step:614/1530 train_loss:3.7011 train_time:101155ms step_avg:167.47ms
step:615/1530 train_loss:3.6978 train_time:101324ms step_avg:167.48ms
step:616/1530 train_loss:3.6303 train_time:101495ms step_avg:167.48ms
step:617/1530 train_loss:3.5558 train_time:101669ms step_avg:167.49ms
step:618/1530 train_loss:3.6912 train_time:101839ms step_avg:167.50ms
step:619/1530 train_loss:3.5515 train_time:102010ms step_avg:167.50ms
step:620/1530 train_loss:3.5913 train_time:102180ms step_avg:167.51ms
step:621/1530 train_loss:3.9289 train_time:102353ms step_avg:167.52ms
step:622/1530 train_loss:3.5720 train_time:102525ms step_avg:167.53ms
step:623/1530 train_loss:3.6060 train_time:102698ms step_avg:167.53ms
step:624/1530 train_loss:3.6946 train_time:102869ms step_avg:167.54ms
step:625/1530 train_loss:3.7008 train_time:103040ms step_avg:167.55ms
step:625/1530 val_loss:3.6242 train_time:103090ms step_avg:167.63ms
step:626/1530 train_loss:3.7411 train_time:103211ms step_avg:167.55ms
step:627/1530 train_loss:3.7147 train_time:103383ms step_avg:167.56ms
step:628/1530 train_loss:3.7630 train_time:103553ms step_avg:167.56ms
step:629/1530 train_loss:3.5881 train_time:103724ms step_avg:167.57ms
step:630/1530 train_loss:3.7277 train_time:103895ms step_avg:167.57ms
step:631/1530 train_loss:3.7378 train_time:104065ms step_avg:167.58ms
step:632/1530 train_loss:3.6464 train_time:104239ms step_avg:167.59ms
step:633/1530 train_loss:3.6047 train_time:104413ms step_avg:167.60ms
step:634/1530 train_loss:3.6951 train_time:104582ms step_avg:167.60ms
step:635/1530 train_loss:3.9538 train_time:104751ms step_avg:167.60ms
step:636/1530 train_loss:3.5479 train_time:104923ms step_avg:167.61ms
step:637/1530 train_loss:3.3557 train_time:105096ms step_avg:167.62ms
step:638/1530 train_loss:3.5914 train_time:105265ms step_avg:167.62ms
step:639/1530 train_loss:3.6343 train_time:105436ms step_avg:167.62ms
step:640/1530 train_loss:3.5645 train_time:105605ms step_avg:167.63ms
step:641/1530 train_loss:3.5827 train_time:105777ms step_avg:167.63ms
step:642/1530 train_loss:3.6281 train_time:105948ms step_avg:167.64ms
step:643/1530 train_loss:3.5984 train_time:106120ms step_avg:167.65ms
step:644/1530 train_loss:3.5615 train_time:106290ms step_avg:167.65ms
step:645/1530 train_loss:3.7748 train_time:106461ms step_avg:167.66ms
step:646/1530 train_loss:3.6690 train_time:106634ms step_avg:167.66ms
step:647/1530 train_loss:3.6650 train_time:106804ms step_avg:167.67ms
step:648/1530 train_loss:3.7142 train_time:106977ms step_avg:167.68ms
step:649/1530 train_loss:3.7645 train_time:107146ms step_avg:167.68ms
step:650/1530 train_loss:3.6239 train_time:107320ms step_avg:167.69ms
step:651/1530 train_loss:3.7682 train_time:107491ms step_avg:167.69ms
step:652/1530 train_loss:3.5816 train_time:107660ms step_avg:167.70ms
step:653/1530 train_loss:3.6668 train_time:107830ms step_avg:167.70ms
step:654/1530 train_loss:3.4292 train_time:108003ms step_avg:167.71ms
step:655/1530 train_loss:3.5793 train_time:108172ms step_avg:167.71ms
step:656/1530 train_loss:3.5752 train_time:108342ms step_avg:167.71ms
step:657/1530 train_loss:3.4961 train_time:108514ms step_avg:167.72ms
step:658/1530 train_loss:3.6931 train_time:108684ms step_avg:167.72ms
step:659/1530 train_loss:3.5849 train_time:108855ms step_avg:167.73ms
step:660/1530 train_loss:3.6887 train_time:109025ms step_avg:167.73ms
step:661/1530 train_loss:3.7528 train_time:109198ms step_avg:167.74ms
step:662/1530 train_loss:3.6688 train_time:109368ms step_avg:167.74ms
step:663/1530 train_loss:3.5560 train_time:109539ms step_avg:167.75ms
step:664/1530 train_loss:3.6142 train_time:109710ms step_avg:167.75ms
step:665/1530 train_loss:3.4927 train_time:109881ms step_avg:167.76ms
step:666/1530 train_loss:3.7808 train_time:110052ms step_avg:167.76ms
step:667/1530 train_loss:3.6063 train_time:110222ms step_avg:167.77ms
step:668/1530 train_loss:3.6456 train_time:110393ms step_avg:167.77ms
step:669/1530 train_loss:3.4882 train_time:110563ms step_avg:167.77ms
step:670/1530 train_loss:3.5981 train_time:110734ms step_avg:167.78ms
step:671/1530 train_loss:3.5592 train_time:110904ms step_avg:167.78ms
step:672/1530 train_loss:3.5676 train_time:111077ms step_avg:167.79ms
step:673/1530 train_loss:3.8543 train_time:111246ms step_avg:167.79ms
step:674/1530 train_loss:3.6265 train_time:111419ms step_avg:167.80ms
step:675/1530 train_loss:3.7155 train_time:111591ms step_avg:167.81ms
step:676/1530 train_loss:3.4937 train_time:111762ms step_avg:167.81ms
step:677/1530 train_loss:3.6009 train_time:111933ms step_avg:167.82ms
step:678/1530 train_loss:3.5557 train_time:112105ms step_avg:167.82ms
step:679/1530 train_loss:3.6759 train_time:112277ms step_avg:167.83ms
step:680/1530 train_loss:3.5846 train_time:112448ms step_avg:167.83ms
step:681/1530 train_loss:3.6166 train_time:112622ms step_avg:167.84ms
step:682/1530 train_loss:3.6619 train_time:112800ms step_avg:167.86ms
step:683/1530 train_loss:3.7370 train_time:112972ms step_avg:167.86ms
step:684/1530 train_loss:3.6482 train_time:113143ms step_avg:167.87ms
step:685/1530 train_loss:3.6869 train_time:113320ms step_avg:167.88ms
step:686/1530 train_loss:3.6391 train_time:113493ms step_avg:167.89ms
step:687/1530 train_loss:3.6653 train_time:113663ms step_avg:167.89ms
step:688/1530 train_loss:3.2247 train_time:113839ms step_avg:167.90ms
step:689/1530 train_loss:3.4095 train_time:114012ms step_avg:167.91ms
step:690/1530 train_loss:3.5393 train_time:114187ms step_avg:167.92ms
step:691/1530 train_loss:3.4109 train_time:114360ms step_avg:167.93ms
step:692/1530 train_loss:3.6267 train_time:114531ms step_avg:167.93ms
step:693/1530 train_loss:3.6493 train_time:114705ms step_avg:167.94ms
step:694/1530 train_loss:3.5552 train_time:114878ms step_avg:167.95ms
step:695/1530 train_loss:3.5374 train_time:115049ms step_avg:167.96ms
step:696/1530 train_loss:3.8556 train_time:115223ms step_avg:167.96ms
step:697/1530 train_loss:3.5909 train_time:115396ms step_avg:167.97ms
step:698/1530 train_loss:3.6454 train_time:115567ms step_avg:167.98ms
step:699/1530 train_loss:3.7736 train_time:115743ms step_avg:167.99ms
step:700/1530 train_loss:3.5719 train_time:115915ms step_avg:167.99ms
step:701/1530 train_loss:3.5435 train_time:116087ms step_avg:168.00ms
step:702/1530 train_loss:3.5132 train_time:116260ms step_avg:168.01ms
step:703/1530 train_loss:3.5001 train_time:116431ms step_avg:168.01ms
step:704/1530 train_loss:3.5725 train_time:116605ms step_avg:168.02ms
step:705/1530 train_loss:3.5589 train_time:116782ms step_avg:168.03ms
step:706/1530 train_loss:3.5837 train_time:116957ms step_avg:168.04ms
step:707/1530 train_loss:3.6471 train_time:117132ms step_avg:168.05ms
step:708/1530 train_loss:3.6047 train_time:117304ms step_avg:168.06ms
step:709/1530 train_loss:3.5819 train_time:117479ms step_avg:168.07ms
step:710/1530 train_loss:3.5430 train_time:117649ms step_avg:168.07ms
step:711/1530 train_loss:3.5924 train_time:117824ms step_avg:168.08ms
step:712/1530 train_loss:3.6507 train_time:118000ms step_avg:168.09ms
step:713/1530 train_loss:3.6547 train_time:118175ms step_avg:168.10ms
step:714/1530 train_loss:3.5619 train_time:118348ms step_avg:168.11ms
step:715/1530 train_loss:3.5689 train_time:118522ms step_avg:168.12ms
step:716/1530 train_loss:3.5881 train_time:118694ms step_avg:168.12ms
step:717/1530 train_loss:3.7101 train_time:118867ms step_avg:168.13ms
step:718/1530 train_loss:3.6018 train_time:119039ms step_avg:168.13ms
step:719/1530 train_loss:3.6794 train_time:119212ms step_avg:168.14ms
step:720/1530 train_loss:3.8478 train_time:119386ms step_avg:168.15ms
step:721/1530 train_loss:3.4663 train_time:119559ms step_avg:168.16ms
step:722/1530 train_loss:3.7407 train_time:119731ms step_avg:168.16ms
step:723/1530 train_loss:3.7738 train_time:119903ms step_avg:168.17ms
step:724/1530 train_loss:3.5736 train_time:120077ms step_avg:168.17ms
step:725/1530 train_loss:3.6527 train_time:120249ms step_avg:168.18ms
step:726/1530 train_loss:3.5273 train_time:120424ms step_avg:168.19ms
step:727/1530 train_loss:3.5759 train_time:120601ms step_avg:168.20ms
step:728/1530 train_loss:3.7292 train_time:120773ms step_avg:168.21ms
step:729/1530 train_loss:3.6682 train_time:120944ms step_avg:168.21ms
step:730/1530 train_loss:3.6602 train_time:121120ms step_avg:168.22ms
step:731/1530 train_loss:3.5527 train_time:121292ms step_avg:168.23ms
step:732/1530 train_loss:3.5976 train_time:121462ms step_avg:168.23ms
step:733/1530 train_loss:3.8312 train_time:121638ms step_avg:168.24ms
step:734/1530 train_loss:3.5633 train_time:121813ms step_avg:168.25ms
step:735/1530 train_loss:3.6140 train_time:121984ms step_avg:168.25ms
step:736/1530 train_loss:3.7365 train_time:122157ms step_avg:168.26ms
step:737/1530 train_loss:3.6780 train_time:122329ms step_avg:168.26ms
step:738/1530 train_loss:3.6011 train_time:122502ms step_avg:168.27ms
step:739/1530 train_loss:3.5028 train_time:122674ms step_avg:168.28ms
step:740/1530 train_loss:4.1124 train_time:122850ms step_avg:168.29ms
step:741/1530 train_loss:3.4927 train_time:123022ms step_avg:168.29ms
step:742/1530 train_loss:3.5632 train_time:123197ms step_avg:168.30ms
step:743/1530 train_loss:3.5824 train_time:123368ms step_avg:168.31ms
step:744/1530 train_loss:3.6468 train_time:123542ms step_avg:168.31ms
step:745/1530 train_loss:3.5822 train_time:123716ms step_avg:168.32ms
step:746/1530 train_loss:3.5990 train_time:123887ms step_avg:168.33ms
step:747/1530 train_loss:3.6491 train_time:124062ms step_avg:168.33ms
step:748/1530 train_loss:3.5681 train_time:124239ms step_avg:168.35ms
step:749/1530 train_loss:3.5619 train_time:124412ms step_avg:168.35ms
step:750/1530 train_loss:3.5970 train_time:124583ms step_avg:168.35ms
step:750/1530 val_loss:3.5666 train_time:124633ms step_avg:168.42ms
step:751/1530 train_loss:3.5700 train_time:124756ms step_avg:168.36ms
step:752/1530 train_loss:3.6197 train_time:124928ms step_avg:168.37ms
step:753/1530 train_loss:3.6178 train_time:125102ms step_avg:168.37ms
step:754/1530 train_loss:3.5914 train_time:125275ms step_avg:168.38ms
step:755/1530 train_loss:3.6795 train_time:125588ms step_avg:168.57ms
step:756/1530 train_loss:3.4604 train_time:125773ms step_avg:168.60ms
step:757/1530 train_loss:3.7309 train_time:125946ms step_avg:168.60ms
step:758/1530 train_loss:3.6517 train_time:126117ms step_avg:168.61ms
step:759/1530 train_loss:3.5928 train_time:126439ms step_avg:168.81ms
step:760/1530 train_loss:3.7051 train_time:126609ms step_avg:168.81ms
step:761/1530 train_loss:3.4055 train_time:126781ms step_avg:168.82ms
step:762/1530 train_loss:3.5486 train_time:126953ms step_avg:168.82ms
step:763/1530 train_loss:3.6590 train_time:127126ms step_avg:168.83ms
step:764/1530 train_loss:3.3187 train_time:127298ms step_avg:168.83ms
step:765/1530 train_loss:3.7309 train_time:127470ms step_avg:168.83ms
step:766/1530 train_loss:3.5725 train_time:127645ms step_avg:168.84ms
step:767/1530 train_loss:3.5706 train_time:127817ms step_avg:168.85ms
step:768/1530 train_loss:3.5708 train_time:127988ms step_avg:168.85ms
step:769/1530 train_loss:3.5853 train_time:128162ms step_avg:168.86ms
step:770/1530 train_loss:3.6374 train_time:128336ms step_avg:168.86ms
step:771/1530 train_loss:3.8812 train_time:128507ms step_avg:168.87ms
step:772/1530 train_loss:3.4490 train_time:128678ms step_avg:168.87ms
step:773/1530 train_loss:3.6306 train_time:128849ms step_avg:168.87ms
step:774/1530 train_loss:3.6394 train_time:129021ms step_avg:168.88ms
step:775/1530 train_loss:3.6112 train_time:129193ms step_avg:168.88ms
step:776/1530 train_loss:3.4091 train_time:129366ms step_avg:168.88ms
step:777/1530 train_loss:3.3841 train_time:129541ms step_avg:168.89ms
step:778/1530 train_loss:3.4920 train_time:129713ms step_avg:168.90ms
step:779/1530 train_loss:3.5788 train_time:129885ms step_avg:168.90ms
step:780/1530 train_loss:3.5825 train_time:130058ms step_avg:168.91ms
step:781/1530 train_loss:3.6779 train_time:130228ms step_avg:168.91ms
step:782/1530 train_loss:3.5864 train_time:130402ms step_avg:168.91ms
step:783/1530 train_loss:3.5670 train_time:130573ms step_avg:168.92ms
step:784/1530 train_loss:3.6047 train_time:130745ms step_avg:168.92ms
step:785/1530 train_loss:3.5624 train_time:130918ms step_avg:168.93ms
step:786/1530 train_loss:3.4413 train_time:131090ms step_avg:168.93ms
step:787/1530 train_loss:3.7175 train_time:131263ms step_avg:168.94ms
step:788/1530 train_loss:3.4973 train_time:131439ms step_avg:168.95ms
step:789/1530 train_loss:3.5479 train_time:131611ms step_avg:168.95ms
step:790/1530 train_loss:3.6284 train_time:131785ms step_avg:168.95ms
step:791/1530 train_loss:3.7727 train_time:131961ms step_avg:168.96ms
step:792/1530 train_loss:3.7541 train_time:132136ms step_avg:168.97ms
step:793/1530 train_loss:3.4576 train_time:132307ms step_avg:168.97ms
step:794/1530 train_loss:3.5905 train_time:132480ms step_avg:168.98ms
step:795/1530 train_loss:3.6757 train_time:132654ms step_avg:168.99ms
step:796/1530 train_loss:3.7424 train_time:132831ms step_avg:169.00ms
step:797/1530 train_loss:3.5215 train_time:133005ms step_avg:169.00ms
step:798/1530 train_loss:3.6482 train_time:133180ms step_avg:169.01ms
step:799/1530 train_loss:3.5437 train_time:133356ms step_avg:169.02ms
step:800/1530 train_loss:3.5344 train_time:133530ms step_avg:169.02ms
step:801/1530 train_loss:3.6249 train_time:133703ms step_avg:169.03ms
step:802/1530 train_loss:3.4931 train_time:133879ms step_avg:169.04ms
step:803/1530 train_loss:3.4875 train_time:134053ms step_avg:169.05ms
step:804/1530 train_loss:3.6215 train_time:134228ms step_avg:169.05ms
step:805/1530 train_loss:3.5137 train_time:134404ms step_avg:169.06ms
step:806/1530 train_loss:3.5595 train_time:134577ms step_avg:169.07ms
step:807/1530 train_loss:3.6415 train_time:134749ms step_avg:169.07ms
step:808/1530 train_loss:3.5421 train_time:134925ms step_avg:169.08ms
step:809/1530 train_loss:3.4924 train_time:135099ms step_avg:169.09ms
step:810/1530 train_loss:3.5630 train_time:135270ms step_avg:169.09ms
step:811/1530 train_loss:3.5814 train_time:135444ms step_avg:169.09ms
step:812/1530 train_loss:3.6002 train_time:135619ms step_avg:169.10ms
step:813/1530 train_loss:3.6247 train_time:135790ms step_avg:169.10ms
step:814/1530 train_loss:3.5659 train_time:135965ms step_avg:169.11ms
step:815/1530 train_loss:3.5676 train_time:136142ms step_avg:169.12ms
step:816/1530 train_loss:3.6822 train_time:136318ms step_avg:169.13ms
step:817/1530 train_loss:3.7676 train_time:136491ms step_avg:169.13ms
step:818/1530 train_loss:3.5212 train_time:136664ms step_avg:169.14ms
step:819/1530 train_loss:3.7194 train_time:136839ms step_avg:169.15ms
step:820/1530 train_loss:3.4956 train_time:137013ms step_avg:169.15ms
step:821/1530 train_loss:3.5641 train_time:137185ms step_avg:169.16ms
step:822/1530 train_loss:3.6947 train_time:137361ms step_avg:169.16ms
step:823/1530 train_loss:3.5772 train_time:137535ms step_avg:169.17ms
step:824/1530 train_loss:3.5107 train_time:137708ms step_avg:169.17ms
step:825/1530 train_loss:3.6161 train_time:137882ms step_avg:169.18ms
step:826/1530 train_loss:3.4791 train_time:138058ms step_avg:169.19ms
step:827/1530 train_loss:3.7329 train_time:138233ms step_avg:169.20ms
step:828/1530 train_loss:3.6204 train_time:138406ms step_avg:169.20ms
step:829/1530 train_loss:3.6235 train_time:138582ms step_avg:169.21ms
step:830/1530 train_loss:3.5315 train_time:138756ms step_avg:169.22ms
step:831/1530 train_loss:3.5999 train_time:138929ms step_avg:169.22ms
step:832/1530 train_loss:3.5195 train_time:139105ms step_avg:169.23ms
step:833/1530 train_loss:3.6518 train_time:139280ms step_avg:169.23ms
step:834/1530 train_loss:3.4717 train_time:139454ms step_avg:169.24ms
step:835/1530 train_loss:3.4549 train_time:139627ms step_avg:169.25ms
step:836/1530 train_loss:3.7135 train_time:139804ms step_avg:169.25ms
step:837/1530 train_loss:3.3949 train_time:139979ms step_avg:169.26ms
step:838/1530 train_loss:3.5950 train_time:140152ms step_avg:169.27ms
step:839/1530 train_loss:3.4226 train_time:140327ms step_avg:169.27ms
step:840/1530 train_loss:3.4684 train_time:140500ms step_avg:169.28ms
step:841/1530 train_loss:3.5640 train_time:140673ms step_avg:169.28ms
step:842/1530 train_loss:3.5871 train_time:140847ms step_avg:169.29ms
step:843/1530 train_loss:3.5582 train_time:141021ms step_avg:169.29ms
step:844/1530 train_loss:3.4269 train_time:141194ms step_avg:169.30ms
step:845/1530 train_loss:3.6571 train_time:141366ms step_avg:169.30ms
step:846/1530 train_loss:3.5152 train_time:141544ms step_avg:169.31ms
step:847/1530 train_loss:3.4896 train_time:141720ms step_avg:169.32ms
step:848/1530 train_loss:3.6395 train_time:141893ms step_avg:169.32ms
step:849/1530 train_loss:3.4929 train_time:142067ms step_avg:169.33ms
step:850/1530 train_loss:3.4360 train_time:142243ms step_avg:169.34ms
step:851/1530 train_loss:3.7371 train_time:142418ms step_avg:169.34ms
step:852/1530 train_loss:3.4448 train_time:142591ms step_avg:169.35ms
step:853/1530 train_loss:3.5651 train_time:142765ms step_avg:169.35ms
step:854/1530 train_loss:3.6487 train_time:142940ms step_avg:169.36ms
step:855/1530 train_loss:3.5139 train_time:143115ms step_avg:169.37ms
step:856/1530 train_loss:3.5445 train_time:143288ms step_avg:169.37ms
step:857/1530 train_loss:3.6013 train_time:143463ms step_avg:169.38ms
step:858/1530 train_loss:3.4586 train_time:143640ms step_avg:169.39ms
step:859/1530 train_loss:3.5593 train_time:143814ms step_avg:169.39ms
step:860/1530 train_loss:3.5843 train_time:143986ms step_avg:169.39ms
step:861/1530 train_loss:3.6235 train_time:144165ms step_avg:169.41ms
step:862/1530 train_loss:3.6025 train_time:144343ms step_avg:169.42ms
step:863/1530 train_loss:3.5682 train_time:144519ms step_avg:169.42ms
step:864/1530 train_loss:3.3804 train_time:144691ms step_avg:169.43ms
step:865/1530 train_loss:3.5955 train_time:144863ms step_avg:169.43ms
step:866/1530 train_loss:3.8891 train_time:145041ms step_avg:169.44ms
step:867/1530 train_loss:3.4590 train_time:145213ms step_avg:169.44ms
step:868/1530 train_loss:3.6426 train_time:145385ms step_avg:169.45ms
step:869/1530 train_loss:3.6168 train_time:145559ms step_avg:169.45ms
step:870/1530 train_loss:3.4463 train_time:145734ms step_avg:169.46ms
step:871/1530 train_loss:3.3901 train_time:145906ms step_avg:169.46ms
step:872/1530 train_loss:3.6533 train_time:146081ms step_avg:169.47ms
step:873/1530 train_loss:3.4597 train_time:146255ms step_avg:169.47ms
step:874/1530 train_loss:3.2236 train_time:146434ms step_avg:169.48ms
step:875/1530 train_loss:3.6347 train_time:146606ms step_avg:169.49ms
step:875/1530 val_loss:3.5200 train_time:146657ms step_avg:169.55ms
step:876/1530 train_loss:3.4393 train_time:146782ms step_avg:169.49ms
step:877/1530 train_loss:3.6228 train_time:146958ms step_avg:169.50ms
step:878/1530 train_loss:3.4716 train_time:147133ms step_avg:169.51ms
step:879/1530 train_loss:3.6465 train_time:147305ms step_avg:169.51ms
step:880/1530 train_loss:3.3110 train_time:147476ms step_avg:169.51ms
step:881/1530 train_loss:3.4789 train_time:147649ms step_avg:169.52ms
step:882/1530 train_loss:3.6974 train_time:147822ms step_avg:169.52ms
step:883/1530 train_loss:3.8428 train_time:147994ms step_avg:169.52ms
step:884/1530 train_loss:3.5693 train_time:148170ms step_avg:169.53ms
step:885/1530 train_loss:3.5021 train_time:148344ms step_avg:169.54ms
step:886/1530 train_loss:3.5730 train_time:148517ms step_avg:169.54ms
step:887/1530 train_loss:4.1024 train_time:148691ms step_avg:169.55ms
step:888/1530 train_loss:3.8309 train_time:148870ms step_avg:169.56ms
step:889/1530 train_loss:3.5186 train_time:149043ms step_avg:169.56ms
step:890/1530 train_loss:3.5323 train_time:149215ms step_avg:169.56ms
step:891/1530 train_loss:3.3597 train_time:149389ms step_avg:169.57ms
step:892/1530 train_loss:3.7186 train_time:149564ms step_avg:169.57ms
step:893/1530 train_loss:3.4247 train_time:149736ms step_avg:169.58ms
step:894/1530 train_loss:3.6383 train_time:149914ms step_avg:169.59ms
step:895/1530 train_loss:3.6808 train_time:150089ms step_avg:169.59ms
step:896/1530 train_loss:3.5027 train_time:150264ms step_avg:169.60ms
step:897/1530 train_loss:3.5391 train_time:150440ms step_avg:169.61ms
step:898/1530 train_loss:3.5842 train_time:150613ms step_avg:169.61ms
step:899/1530 train_loss:3.4764 train_time:150785ms step_avg:169.61ms
step:900/1530 train_loss:3.4307 train_time:150958ms step_avg:169.62ms
step:901/1530 train_loss:3.6239 train_time:151130ms step_avg:169.62ms
step:902/1530 train_loss:3.6359 train_time:151303ms step_avg:169.62ms
step:903/1530 train_loss:3.5435 train_time:151479ms step_avg:169.63ms
step:904/1530 train_loss:3.4939 train_time:151652ms step_avg:169.63ms
step:905/1530 train_loss:3.5008 train_time:151823ms step_avg:169.63ms
step:906/1530 train_loss:3.7041 train_time:151997ms step_avg:169.64ms
step:907/1530 train_loss:3.5198 train_time:152170ms step_avg:169.64ms
step:908/1530 train_loss:3.5706 train_time:152343ms step_avg:169.65ms
step:909/1530 train_loss:3.4558 train_time:152518ms step_avg:169.65ms
step:910/1530 train_loss:3.5347 train_time:152697ms step_avg:169.66ms
step:911/1530 train_loss:3.6464 train_time:152872ms step_avg:169.67ms
step:912/1530 train_loss:3.5928 train_time:153051ms step_avg:169.68ms
step:913/1530 train_loss:3.4652 train_time:153229ms step_avg:169.69ms
step:914/1530 train_loss:3.7448 train_time:153406ms step_avg:169.70ms
step:915/1530 train_loss:3.5327 train_time:153586ms step_avg:169.71ms
step:916/1530 train_loss:3.6181 train_time:153763ms step_avg:169.72ms
step:917/1530 train_loss:3.5999 train_time:153937ms step_avg:169.72ms
step:918/1530 train_loss:4.8316 train_time:154115ms step_avg:169.73ms
step:919/1530 train_loss:3.5012 train_time:154292ms step_avg:169.74ms
step:920/1530 train_loss:3.5915 train_time:154467ms step_avg:169.74ms
step:921/1530 train_loss:3.5501 train_time:154643ms step_avg:169.75ms
step:922/1530 train_loss:3.5857 train_time:154819ms step_avg:169.76ms
step:923/1530 train_loss:3.6111 train_time:154994ms step_avg:169.76ms
step:924/1530 train_loss:3.6791 train_time:155171ms step_avg:169.77ms
step:925/1530 train_loss:3.6471 train_time:155346ms step_avg:169.78ms
step:926/1530 train_loss:3.5498 train_time:155520ms step_avg:169.78ms
step:927/1530 train_loss:3.5554 train_time:155695ms step_avg:169.79ms
step:928/1530 train_loss:3.7725 train_time:155872ms step_avg:169.80ms
step:929/1530 train_loss:3.6087 train_time:156047ms step_avg:169.80ms
step:930/1530 train_loss:3.4032 train_time:156225ms step_avg:169.81ms
step:931/1530 train_loss:3.4979 train_time:156399ms step_avg:169.81ms
step:932/1530 train_loss:3.6471 train_time:156576ms step_avg:169.82ms
step:933/1530 train_loss:3.3699 train_time:156753ms step_avg:169.83ms
step:934/1530 train_loss:3.5868 train_time:156932ms step_avg:169.84ms
step:935/1530 train_loss:3.4383 train_time:157110ms step_avg:169.85ms
step:936/1530 train_loss:3.5226 train_time:157289ms step_avg:169.86ms
step:937/1530 train_loss:3.6255 train_time:157468ms step_avg:169.87ms
step:938/1530 train_loss:3.5408 train_time:157643ms step_avg:169.87ms
step:939/1530 train_loss:3.6675 train_time:157823ms step_avg:169.88ms
step:940/1530 train_loss:3.4819 train_time:157998ms step_avg:169.89ms
step:941/1530 train_loss:3.5514 train_time:158172ms step_avg:169.90ms
step:942/1530 train_loss:3.3576 train_time:158351ms step_avg:169.90ms
step:943/1530 train_loss:3.7140 train_time:158532ms step_avg:169.92ms
step:944/1530 train_loss:3.4006 train_time:158849ms step_avg:170.07ms
step:945/1530 train_loss:3.4256 train_time:159033ms step_avg:170.09ms
step:946/1530 train_loss:5.0715 train_time:159214ms step_avg:170.10ms
step:947/1530 train_loss:3.5985 train_time:159390ms step_avg:170.11ms
step:948/1530 train_loss:3.4889 train_time:159564ms step_avg:170.11ms
step:949/1530 train_loss:3.3763 train_time:159887ms step_avg:170.27ms
step:950/1530 train_loss:3.4445 train_time:160061ms step_avg:170.28ms
step:951/1530 train_loss:3.4075 train_time:160238ms step_avg:170.28ms
step:952/1530 train_loss:3.4819 train_time:160414ms step_avg:170.29ms
step:953/1530 train_loss:3.5702 train_time:160592ms step_avg:170.30ms
step:954/1530 train_loss:3.4454 train_time:160772ms step_avg:170.31ms
step:955/1530 train_loss:3.4775 train_time:160947ms step_avg:170.31ms
step:956/1530 train_loss:3.4467 train_time:161123ms step_avg:170.32ms
step:957/1530 train_loss:3.4978 train_time:161301ms step_avg:170.33ms
step:958/1530 train_loss:3.5028 train_time:161481ms step_avg:170.34ms
step:959/1530 train_loss:3.5158 train_time:161657ms step_avg:170.34ms
step:960/1530 train_loss:3.4102 train_time:161834ms step_avg:170.35ms
step:961/1530 train_loss:3.6433 train_time:162009ms step_avg:170.36ms
step:962/1530 train_loss:3.5929 train_time:162186ms step_avg:170.36ms
step:963/1530 train_loss:3.4894 train_time:162366ms step_avg:170.37ms
step:964/1530 train_loss:3.4291 train_time:162546ms step_avg:170.38ms
step:965/1530 train_loss:3.4771 train_time:162719ms step_avg:170.39ms
step:966/1530 train_loss:3.7057 train_time:162893ms step_avg:170.39ms
step:967/1530 train_loss:3.5197 train_time:163070ms step_avg:170.40ms
step:968/1530 train_loss:3.5153 train_time:163248ms step_avg:170.40ms
step:969/1530 train_loss:3.5857 train_time:163423ms step_avg:170.41ms
step:970/1530 train_loss:3.3787 train_time:163595ms step_avg:170.41ms
step:971/1530 train_loss:3.5350 train_time:163769ms step_avg:170.42ms
step:972/1530 train_loss:3.4762 train_time:163944ms step_avg:170.42ms
step:973/1530 train_loss:3.5393 train_time:164116ms step_avg:170.42ms
step:974/1530 train_loss:3.5900 train_time:164293ms step_avg:170.43ms
step:975/1530 train_loss:3.4649 train_time:164468ms step_avg:170.43ms
step:976/1530 train_loss:3.6706 train_time:164644ms step_avg:170.44ms
step:977/1530 train_loss:3.5713 train_time:164818ms step_avg:170.44ms
step:978/1530 train_loss:3.3568 train_time:164993ms step_avg:170.45ms
step:979/1530 train_loss:3.6233 train_time:165167ms step_avg:170.45ms
step:980/1530 train_loss:3.4176 train_time:165346ms step_avg:170.46ms
step:981/1530 train_loss:3.5783 train_time:165523ms step_avg:170.47ms
step:982/1530 train_loss:3.5396 train_time:165697ms step_avg:170.47ms
step:983/1530 train_loss:3.5137 train_time:165873ms step_avg:170.48ms
step:984/1530 train_loss:3.4940 train_time:166049ms step_avg:170.48ms
step:985/1530 train_loss:3.5736 train_time:166228ms step_avg:170.49ms
step:986/1530 train_loss:3.4122 train_time:166403ms step_avg:170.50ms
step:987/1530 train_loss:3.4902 train_time:166576ms step_avg:170.50ms
step:988/1530 train_loss:3.4719 train_time:166750ms step_avg:170.50ms
step:989/1530 train_loss:3.4167 train_time:166924ms step_avg:170.50ms
step:990/1530 train_loss:3.6551 train_time:167101ms step_avg:170.51ms
step:991/1530 train_loss:3.4677 train_time:167275ms step_avg:170.52ms
step:992/1530 train_loss:3.4438 train_time:167455ms step_avg:170.52ms
step:993/1530 train_loss:3.5015 train_time:167637ms step_avg:170.54ms
step:994/1530 train_loss:3.5975 train_time:167813ms step_avg:170.54ms
step:995/1530 train_loss:3.5334 train_time:167986ms step_avg:170.54ms
step:996/1530 train_loss:3.4583 train_time:168160ms step_avg:170.55ms
step:997/1530 train_loss:3.7586 train_time:168333ms step_avg:170.55ms
step:998/1530 train_loss:3.4386 train_time:168506ms step_avg:170.55ms
step:999/1530 train_loss:3.5885 train_time:168682ms step_avg:170.56ms
step:1000/1530 train_loss:3.4398 train_time:168861ms step_avg:170.57ms
step:1000/1530 val_loss:3.4663 train_time:168910ms step_avg:170.62ms
step:1001/1530 train_loss:3.5006 train_time:169036ms step_avg:170.57ms
step:1002/1530 train_loss:3.3784 train_time:169209ms step_avg:170.57ms
step:1003/1530 train_loss:3.5535 train_time:169385ms step_avg:170.58ms
step:1004/1530 train_loss:3.6055 train_time:169562ms step_avg:170.59ms
step:1005/1530 train_loss:3.3944 train_time:169737ms step_avg:170.59ms
step:1006/1530 train_loss:3.4621 train_time:169913ms step_avg:170.60ms
step:1007/1530 train_loss:3.4410 train_time:170087ms step_avg:170.60ms
step:1008/1530 train_loss:3.5599 train_time:170265ms step_avg:170.61ms
step:1009/1530 train_loss:3.6651 train_time:170443ms step_avg:170.61ms
step:1010/1530 train_loss:3.5540 train_time:170617ms step_avg:170.62ms
step:1011/1530 train_loss:3.5317 train_time:170789ms step_avg:170.62ms
step:1012/1530 train_loss:3.3917 train_time:170964ms step_avg:170.62ms
step:1013/1530 train_loss:3.5337 train_time:171140ms step_avg:170.63ms
step:1014/1530 train_loss:3.6187 train_time:171317ms step_avg:170.63ms
step:1015/1530 train_loss:3.3324 train_time:171494ms step_avg:170.64ms
step:1016/1530 train_loss:3.4081 train_time:171668ms step_avg:170.64ms
step:1017/1530 train_loss:3.4006 train_time:171846ms step_avg:170.65ms
step:1018/1530 train_loss:3.3940 train_time:172021ms step_avg:170.66ms
step:1019/1530 train_loss:3.5227 train_time:172198ms step_avg:170.66ms
step:1020/1530 train_loss:3.3841 train_time:172376ms step_avg:170.67ms
step:1021/1530 train_loss:3.3502 train_time:172549ms step_avg:170.67ms
step:1022/1530 train_loss:3.4806 train_time:172726ms step_avg:170.68ms
step:1023/1530 train_loss:3.5048 train_time:172903ms step_avg:170.68ms
step:1024/1530 train_loss:3.4775 train_time:173081ms step_avg:170.69ms
step:1025/1530 train_loss:3.4766 train_time:173260ms step_avg:170.70ms
step:1026/1530 train_loss:3.6149 train_time:173436ms step_avg:170.70ms
step:1027/1530 train_loss:3.3205 train_time:173611ms step_avg:170.71ms
step:1028/1530 train_loss:3.3950 train_time:173791ms step_avg:170.72ms
step:1029/1530 train_loss:3.3136 train_time:173973ms step_avg:170.73ms
step:1030/1530 train_loss:3.5421 train_time:174148ms step_avg:170.73ms
step:1031/1530 train_loss:3.5069 train_time:174325ms step_avg:170.74ms
step:1032/1530 train_loss:3.6966 train_time:174505ms step_avg:170.75ms
step:1033/1530 train_loss:3.4874 train_time:174683ms step_avg:170.76ms
step:1034/1530 train_loss:3.3956 train_time:174859ms step_avg:170.76ms
step:1035/1530 train_loss:3.4431 train_time:175036ms step_avg:170.77ms
step:1036/1530 train_loss:3.4828 train_time:175211ms step_avg:170.77ms
step:1037/1530 train_loss:3.7865 train_time:175389ms step_avg:170.78ms
step:1038/1530 train_loss:3.6204 train_time:175568ms step_avg:170.79ms
step:1039/1530 train_loss:3.5071 train_time:175748ms step_avg:170.80ms
step:1040/1530 train_loss:3.4166 train_time:175924ms step_avg:170.80ms
step:1041/1530 train_loss:3.4896 train_time:176104ms step_avg:170.81ms
step:1042/1530 train_loss:3.5219 train_time:176277ms step_avg:170.81ms
step:1043/1530 train_loss:3.4447 train_time:176452ms step_avg:170.81ms
step:1044/1530 train_loss:3.4567 train_time:176627ms step_avg:170.82ms
step:1045/1530 train_loss:3.5147 train_time:176805ms step_avg:170.83ms
step:1046/1530 train_loss:3.4238 train_time:176982ms step_avg:170.83ms
step:1047/1530 train_loss:3.6329 train_time:177159ms step_avg:170.84ms
step:1048/1530 train_loss:3.4999 train_time:177334ms step_avg:170.84ms
step:1049/1530 train_loss:3.4012 train_time:177508ms step_avg:170.85ms
step:1050/1530 train_loss:3.3932 train_time:177687ms step_avg:170.85ms
step:1051/1530 train_loss:3.4957 train_time:177864ms step_avg:170.86ms
step:1052/1530 train_loss:3.3614 train_time:178043ms step_avg:170.87ms
step:1053/1530 train_loss:3.6861 train_time:178220ms step_avg:170.87ms
step:1054/1530 train_loss:3.5392 train_time:178399ms step_avg:170.88ms
step:1055/1530 train_loss:3.3858 train_time:178575ms step_avg:170.89ms
step:1056/1530 train_loss:3.4999 train_time:178750ms step_avg:170.89ms
step:1057/1530 train_loss:3.5778 train_time:178927ms step_avg:170.90ms
step:1058/1530 train_loss:3.3031 train_time:179106ms step_avg:170.90ms
step:1059/1530 train_loss:3.3709 train_time:179287ms step_avg:170.91ms
step:1060/1530 train_loss:3.4355 train_time:179464ms step_avg:170.92ms
step:1061/1530 train_loss:3.4163 train_time:179638ms step_avg:170.92ms
step:1062/1530 train_loss:3.3800 train_time:179815ms step_avg:170.93ms
step:1063/1530 train_loss:3.4606 train_time:179988ms step_avg:170.93ms
step:1064/1530 train_loss:3.3812 train_time:180163ms step_avg:170.93ms
step:1065/1530 train_loss:3.3599 train_time:180342ms step_avg:170.94ms
step:1066/1530 train_loss:3.4150 train_time:180517ms step_avg:170.94ms
step:1067/1530 train_loss:3.2867 train_time:180696ms step_avg:170.95ms
step:1068/1530 train_loss:3.4372 train_time:180871ms step_avg:170.96ms
step:1069/1530 train_loss:3.2978 train_time:181052ms step_avg:170.96ms
step:1070/1530 train_loss:3.5683 train_time:181226ms step_avg:170.97ms
step:1071/1530 train_loss:3.5141 train_time:181404ms step_avg:170.97ms
step:1072/1530 train_loss:3.4411 train_time:181580ms step_avg:170.98ms
step:1073/1530 train_loss:3.5200 train_time:181753ms step_avg:170.98ms
step:1074/1530 train_loss:3.4309 train_time:181930ms step_avg:170.99ms
step:1075/1530 train_loss:3.4027 train_time:182108ms step_avg:170.99ms
step:1076/1530 train_loss:3.7946 train_time:182285ms step_avg:171.00ms
step:1077/1530 train_loss:3.4333 train_time:182461ms step_avg:171.00ms
step:1078/1530 train_loss:3.0957 train_time:182646ms step_avg:171.02ms
step:1079/1530 train_loss:3.5350 train_time:182822ms step_avg:171.02ms
step:1080/1530 train_loss:3.4312 train_time:183001ms step_avg:171.03ms
step:1081/1530 train_loss:3.5035 train_time:183177ms step_avg:171.03ms
step:1082/1530 train_loss:3.5845 train_time:183352ms step_avg:171.04ms
step:1083/1530 train_loss:3.4978 train_time:183526ms step_avg:171.04ms
step:1084/1530 train_loss:3.4645 train_time:183704ms step_avg:171.05ms
step:1085/1530 train_loss:3.4329 train_time:183880ms step_avg:171.05ms
step:1086/1530 train_loss:3.6254 train_time:184057ms step_avg:171.06ms
step:1087/1530 train_loss:3.5034 train_time:184230ms step_avg:171.06ms
step:1088/1530 train_loss:3.3669 train_time:184408ms step_avg:171.07ms
step:1089/1530 train_loss:3.3784 train_time:184587ms step_avg:171.07ms
step:1090/1530 train_loss:3.4820 train_time:184766ms step_avg:171.08ms
step:1091/1530 train_loss:3.2844 train_time:184944ms step_avg:171.09ms
step:1092/1530 train_loss:3.4843 train_time:185120ms step_avg:171.09ms
step:1093/1530 train_loss:3.6024 train_time:185298ms step_avg:171.10ms
step:1094/1530 train_loss:3.4473 train_time:185473ms step_avg:171.10ms
step:1095/1530 train_loss:3.4181 train_time:185647ms step_avg:171.10ms
step:1096/1530 train_loss:3.4251 train_time:185824ms step_avg:171.11ms
step:1097/1530 train_loss:3.4906 train_time:186002ms step_avg:171.11ms
step:1098/1530 train_loss:3.5619 train_time:186181ms step_avg:171.12ms
step:1099/1530 train_loss:3.5263 train_time:186356ms step_avg:171.13ms
step:1100/1530 train_loss:3.4299 train_time:186533ms step_avg:171.13ms
step:1101/1530 train_loss:3.2894 train_time:186709ms step_avg:171.14ms
step:1102/1530 train_loss:3.3107 train_time:186888ms step_avg:171.14ms
step:1103/1530 train_loss:3.4440 train_time:187071ms step_avg:171.15ms
step:1104/1530 train_loss:3.3238 train_time:187247ms step_avg:171.16ms
step:1105/1530 train_loss:4.0635 train_time:187426ms step_avg:171.17ms
step:1106/1530 train_loss:3.2255 train_time:187601ms step_avg:171.17ms
step:1107/1530 train_loss:3.5652 train_time:187777ms step_avg:171.17ms
step:1108/1530 train_loss:3.3483 train_time:187952ms step_avg:171.18ms
step:1109/1530 train_loss:3.4984 train_time:188126ms step_avg:171.18ms
step:1110/1530 train_loss:3.4289 train_time:188301ms step_avg:171.18ms
step:1111/1530 train_loss:3.4823 train_time:188477ms step_avg:171.19ms
step:1112/1530 train_loss:3.5559 train_time:188656ms step_avg:171.19ms
step:1113/1530 train_loss:3.4337 train_time:188839ms step_avg:171.20ms
step:1114/1530 train_loss:3.3697 train_time:189018ms step_avg:171.21ms
step:1115/1530 train_loss:3.2387 train_time:189197ms step_avg:171.22ms
step:1116/1530 train_loss:3.4282 train_time:189370ms step_avg:171.22ms
step:1117/1530 train_loss:3.5935 train_time:189548ms step_avg:171.23ms
step:1118/1530 train_loss:3.6198 train_time:189726ms step_avg:171.23ms
step:1119/1530 train_loss:3.4812 train_time:189901ms step_avg:171.24ms
step:1120/1530 train_loss:3.4914 train_time:190081ms step_avg:171.24ms
step:1121/1530 train_loss:3.3930 train_time:190258ms step_avg:171.25ms
step:1122/1530 train_loss:3.4606 train_time:190432ms step_avg:171.25ms
step:1123/1530 train_loss:3.5793 train_time:190609ms step_avg:171.26ms
step:1124/1530 train_loss:3.3397 train_time:190784ms step_avg:171.26ms
step:1125/1530 train_loss:3.2280 train_time:190962ms step_avg:171.27ms
step:1125/1530 val_loss:3.4088 train_time:191012ms step_avg:171.31ms
step:1126/1530 train_loss:3.4703 train_time:191139ms step_avg:171.27ms
step:1127/1530 train_loss:3.6747 train_time:191320ms step_avg:171.28ms
step:1128/1530 train_loss:3.2315 train_time:191498ms step_avg:171.29ms
step:1129/1530 train_loss:3.5542 train_time:191679ms step_avg:171.29ms
step:1130/1530 train_loss:3.3808 train_time:191856ms step_avg:171.30ms
step:1131/1530 train_loss:3.4022 train_time:192035ms step_avg:171.31ms
step:1132/1530 train_loss:3.3669 train_time:192210ms step_avg:171.31ms
step:1133/1530 train_loss:3.4908 train_time:192522ms step_avg:171.44ms
step:1134/1530 train_loss:3.4452 train_time:192707ms step_avg:171.45ms
step:1135/1530 train_loss:3.5221 train_time:192884ms step_avg:171.45ms
step:1136/1530 train_loss:3.5599 train_time:193062ms step_avg:171.46ms
step:1137/1530 train_loss:3.4609 train_time:193237ms step_avg:171.46ms
step:1138/1530 train_loss:3.3517 train_time:193417ms step_avg:171.47ms
step:1139/1530 train_loss:3.6550 train_time:193743ms step_avg:171.61ms
step:1140/1530 train_loss:3.4556 train_time:193920ms step_avg:171.61ms
step:1141/1530 train_loss:3.5953 train_time:194101ms step_avg:171.62ms
step:1142/1530 train_loss:3.4419 train_time:194280ms step_avg:171.63ms
step:1143/1530 train_loss:3.3630 train_time:194458ms step_avg:171.63ms
step:1144/1530 train_loss:3.4447 train_time:194635ms step_avg:171.64ms
step:1145/1530 train_loss:3.5912 train_time:194810ms step_avg:171.64ms
step:1146/1530 train_loss:3.5556 train_time:194990ms step_avg:171.65ms
step:1147/1530 train_loss:3.4851 train_time:195170ms step_avg:171.65ms
step:1148/1530 train_loss:3.5001 train_time:195346ms step_avg:171.66ms
step:1149/1530 train_loss:3.3256 train_time:195526ms step_avg:171.67ms
step:1150/1530 train_loss:3.3746 train_time:195702ms step_avg:171.67ms
step:1151/1530 train_loss:3.3242 train_time:195881ms step_avg:171.67ms
step:1152/1530 train_loss:3.3952 train_time:196063ms step_avg:171.68ms
step:1153/1530 train_loss:3.4315 train_time:196242ms step_avg:171.69ms
step:1154/1530 train_loss:3.5146 train_time:196418ms step_avg:171.69ms
step:1155/1530 train_loss:3.3218 train_time:196601ms step_avg:171.70ms
step:1156/1530 train_loss:3.5354 train_time:196786ms step_avg:171.72ms
step:1157/1530 train_loss:3.4975 train_time:196964ms step_avg:171.72ms
step:1158/1530 train_loss:3.2505 train_time:197140ms step_avg:171.73ms
step:1159/1530 train_loss:3.3466 train_time:197317ms step_avg:171.73ms
step:1160/1530 train_loss:3.3345 train_time:197491ms step_avg:171.73ms
step:1161/1530 train_loss:3.0772 train_time:197673ms step_avg:171.74ms
step:1162/1530 train_loss:3.4223 train_time:197851ms step_avg:171.75ms
step:1163/1530 train_loss:3.3880 train_time:198031ms step_avg:171.75ms
step:1164/1530 train_loss:3.2938 train_time:198207ms step_avg:171.76ms
step:1165/1530 train_loss:3.2525 train_time:198383ms step_avg:171.76ms
step:1166/1530 train_loss:3.3884 train_time:198561ms step_avg:171.77ms
step:1167/1530 train_loss:3.4109 train_time:198736ms step_avg:171.77ms
step:1168/1530 train_loss:3.7236 train_time:198912ms step_avg:171.77ms
step:1169/1530 train_loss:3.3767 train_time:199091ms step_avg:171.78ms
step:1170/1530 train_loss:3.3871 train_time:199266ms step_avg:171.78ms
step:1171/1530 train_loss:3.3146 train_time:199442ms step_avg:171.78ms
step:1172/1530 train_loss:3.4245 train_time:199617ms step_avg:171.79ms
step:1173/1530 train_loss:3.5401 train_time:199797ms step_avg:171.79ms
step:1174/1530 train_loss:3.3812 train_time:199981ms step_avg:171.81ms
step:1175/1530 train_loss:3.3649 train_time:200160ms step_avg:171.81ms
step:1176/1530 train_loss:3.4248 train_time:200341ms step_avg:171.82ms
step:1177/1530 train_loss:3.4494 train_time:200522ms step_avg:171.83ms
step:1178/1530 train_loss:3.4973 train_time:200700ms step_avg:171.83ms
step:1179/1530 train_loss:3.3993 train_time:200877ms step_avg:171.84ms
step:1180/1530 train_loss:3.3499 train_time:201064ms step_avg:171.85ms
step:1181/1530 train_loss:3.3363 train_time:201242ms step_avg:171.85ms
step:1182/1530 train_loss:3.3762 train_time:201419ms step_avg:171.86ms
step:1183/1530 train_loss:3.3320 train_time:201598ms step_avg:171.87ms
step:1184/1530 train_loss:3.5114 train_time:201776ms step_avg:171.87ms
step:1185/1530 train_loss:3.5455 train_time:201958ms step_avg:171.88ms
step:1186/1530 train_loss:3.3619 train_time:202137ms step_avg:171.89ms
step:1187/1530 train_loss:3.4174 train_time:202322ms step_avg:171.90ms
step:1188/1530 train_loss:3.4396 train_time:202498ms step_avg:171.90ms
step:1189/1530 train_loss:3.2764 train_time:202679ms step_avg:171.91ms
step:1190/1530 train_loss:3.4432 train_time:202858ms step_avg:171.91ms
step:1191/1530 train_loss:3.5820 train_time:203037ms step_avg:171.92ms
step:1192/1530 train_loss:3.3973 train_time:203213ms step_avg:171.92ms
step:1193/1530 train_loss:3.2768 train_time:203388ms step_avg:171.93ms
step:1194/1530 train_loss:3.5584 train_time:203566ms step_avg:171.93ms
step:1195/1530 train_loss:3.3738 train_time:203746ms step_avg:171.94ms
step:1196/1530 train_loss:3.3860 train_time:203932ms step_avg:171.95ms
step:1197/1530 train_loss:3.2951 train_time:204112ms step_avg:171.96ms
step:1198/1530 train_loss:3.3024 train_time:204297ms step_avg:171.97ms
step:1199/1530 train_loss:3.3435 train_time:204477ms step_avg:171.97ms
step:1200/1530 train_loss:3.4460 train_time:204653ms step_avg:171.98ms
step:1201/1530 train_loss:3.4836 train_time:204831ms step_avg:171.98ms
step:1202/1530 train_loss:3.6046 train_time:205018ms step_avg:172.00ms
step:1203/1530 train_loss:3.4054 train_time:205197ms step_avg:172.00ms
step:1204/1530 train_loss:3.3085 train_time:205378ms step_avg:172.01ms
step:1205/1530 train_loss:3.4413 train_time:205555ms step_avg:172.01ms
step:1206/1530 train_loss:3.4739 train_time:205733ms step_avg:172.02ms
step:1207/1530 train_loss:3.5185 train_time:205911ms step_avg:172.02ms
step:1208/1530 train_loss:3.3953 train_time:206086ms step_avg:172.02ms
step:1209/1530 train_loss:3.2434 train_time:206263ms step_avg:172.03ms
step:1210/1530 train_loss:3.3014 train_time:206442ms step_avg:172.04ms
step:1211/1530 train_loss:3.3926 train_time:206621ms step_avg:172.04ms
step:1212/1530 train_loss:3.4009 train_time:206798ms step_avg:172.05ms
step:1213/1530 train_loss:3.4094 train_time:206977ms step_avg:172.05ms
step:1214/1530 train_loss:3.2534 train_time:207158ms step_avg:172.06ms
step:1215/1530 train_loss:3.3970 train_time:207333ms step_avg:172.06ms
step:1216/1530 train_loss:3.3283 train_time:207511ms step_avg:172.07ms
step:1217/1530 train_loss:3.3194 train_time:207690ms step_avg:172.07ms
step:1218/1530 train_loss:3.4056 train_time:207866ms step_avg:172.07ms
step:1219/1530 train_loss:3.2587 train_time:208051ms step_avg:172.08ms
step:1220/1530 train_loss:3.4757 train_time:208225ms step_avg:172.09ms
step:1221/1530 train_loss:3.5045 train_time:208401ms step_avg:172.09ms
step:1222/1530 train_loss:3.4325 train_time:208576ms step_avg:172.09ms
step:1223/1530 train_loss:3.2925 train_time:208754ms step_avg:172.10ms
step:1224/1530 train_loss:3.2567 train_time:208935ms step_avg:172.10ms
step:1225/1530 train_loss:3.3669 train_time:209112ms step_avg:172.11ms
step:1226/1530 train_loss:3.3357 train_time:209293ms step_avg:172.12ms
step:1227/1530 train_loss:3.2778 train_time:209471ms step_avg:172.12ms
step:1228/1530 train_loss:3.4456 train_time:209648ms step_avg:172.12ms
step:1229/1530 train_loss:3.3695 train_time:209826ms step_avg:172.13ms
step:1230/1530 train_loss:3.3974 train_time:210008ms step_avg:172.14ms
step:1231/1530 train_loss:3.5782 train_time:210190ms step_avg:172.15ms
step:1232/1530 train_loss:3.5012 train_time:210369ms step_avg:172.15ms
step:1233/1530 train_loss:3.4239 train_time:210544ms step_avg:172.15ms
step:1234/1530 train_loss:3.5887 train_time:210722ms step_avg:172.16ms
step:1235/1530 train_loss:3.3225 train_time:210905ms step_avg:172.17ms
step:1236/1530 train_loss:3.2900 train_time:211080ms step_avg:172.17ms
step:1237/1530 train_loss:3.2753 train_time:211259ms step_avg:172.17ms
step:1238/1530 train_loss:3.2825 train_time:211442ms step_avg:172.18ms
step:1239/1530 train_loss:3.3324 train_time:211622ms step_avg:172.19ms
step:1240/1530 train_loss:3.3852 train_time:211799ms step_avg:172.19ms
step:1241/1530 train_loss:3.4281 train_time:211979ms step_avg:172.20ms
step:1242/1530 train_loss:3.2947 train_time:212156ms step_avg:172.20ms
step:1243/1530 train_loss:3.4051 train_time:212336ms step_avg:172.21ms
step:1244/1530 train_loss:3.4057 train_time:212511ms step_avg:172.21ms
step:1245/1530 train_loss:3.4096 train_time:212687ms step_avg:172.22ms
step:1246/1530 train_loss:3.2455 train_time:212866ms step_avg:172.22ms
step:1247/1530 train_loss:3.3752 train_time:213040ms step_avg:172.22ms
step:1248/1530 train_loss:3.4298 train_time:213219ms step_avg:172.23ms
step:1249/1530 train_loss:3.4259 train_time:213400ms step_avg:172.24ms
step:1250/1530 train_loss:3.3064 train_time:213580ms step_avg:172.24ms
step:1250/1530 val_loss:3.3554 train_time:213635ms step_avg:172.29ms
step:1251/1530 train_loss:3.4927 train_time:213771ms step_avg:172.26ms
step:1252/1530 train_loss:3.3616 train_time:213947ms step_avg:172.26ms
step:1253/1530 train_loss:3.3127 train_time:214122ms step_avg:172.26ms
step:1254/1530 train_loss:3.4196 train_time:214304ms step_avg:172.27ms
step:1255/1530 train_loss:3.5178 train_time:214494ms step_avg:172.28ms
step:1256/1530 train_loss:3.3104 train_time:214676ms step_avg:172.29ms
step:1257/1530 train_loss:3.3759 train_time:214855ms step_avg:172.30ms
step:1258/1530 train_loss:3.3653 train_time:215039ms step_avg:172.31ms
step:1259/1530 train_loss:3.3257 train_time:215218ms step_avg:172.31ms
step:1260/1530 train_loss:3.2132 train_time:215397ms step_avg:172.32ms
step:1261/1530 train_loss:3.3035 train_time:215577ms step_avg:172.32ms
step:1262/1530 train_loss:3.3262 train_time:215759ms step_avg:172.33ms
step:1263/1530 train_loss:3.2438 train_time:215939ms step_avg:172.34ms
step:1264/1530 train_loss:3.4468 train_time:216115ms step_avg:172.34ms
step:1265/1530 train_loss:3.4257 train_time:216292ms step_avg:172.34ms
step:1266/1530 train_loss:3.4440 train_time:216471ms step_avg:172.35ms
step:1267/1530 train_loss:3.3721 train_time:216651ms step_avg:172.36ms
step:1268/1530 train_loss:3.4074 train_time:216833ms step_avg:172.36ms
step:1269/1530 train_loss:3.2522 train_time:217018ms step_avg:172.37ms
step:1270/1530 train_loss:3.1085 train_time:217194ms step_avg:172.38ms
step:1271/1530 train_loss:3.4054 train_time:217374ms step_avg:172.38ms
step:1272/1530 train_loss:3.3533 train_time:217550ms step_avg:172.38ms
step:1273/1530 train_loss:3.3819 train_time:217732ms step_avg:172.39ms
step:1274/1530 train_loss:3.3600 train_time:217913ms step_avg:172.40ms
step:1275/1530 train_loss:3.4346 train_time:218090ms step_avg:172.40ms
step:1276/1530 train_loss:3.4680 train_time:218264ms step_avg:172.40ms
step:1277/1530 train_loss:3.4159 train_time:218442ms step_avg:172.41ms
step:1278/1530 train_loss:3.4086 train_time:218618ms step_avg:172.41ms
step:1279/1530 train_loss:3.2712 train_time:218799ms step_avg:172.42ms
step:1280/1530 train_loss:3.3720 train_time:218985ms step_avg:172.43ms
step:1281/1530 train_loss:3.4246 train_time:219161ms step_avg:172.43ms
step:1282/1530 train_loss:3.4693 train_time:219336ms step_avg:172.43ms
step:1283/1530 train_loss:3.3307 train_time:219514ms step_avg:172.44ms
step:1284/1530 train_loss:3.3665 train_time:219694ms step_avg:172.44ms
step:1285/1530 train_loss:3.3623 train_time:219874ms step_avg:172.45ms
step:1286/1530 train_loss:3.3347 train_time:220051ms step_avg:172.45ms
step:1287/1530 train_loss:3.4832 train_time:220230ms step_avg:172.46ms
step:1288/1530 train_loss:3.3020 train_time:220412ms step_avg:172.47ms
step:1289/1530 train_loss:3.3828 train_time:220596ms step_avg:172.48ms
step:1290/1530 train_loss:3.4614 train_time:220781ms step_avg:172.48ms
step:1291/1530 train_loss:3.3879 train_time:220962ms step_avg:172.49ms
step:1292/1530 train_loss:3.4778 train_time:221145ms step_avg:172.50ms
step:1293/1530 train_loss:3.5214 train_time:221328ms step_avg:172.51ms
step:1294/1530 train_loss:3.4656 train_time:221508ms step_avg:172.51ms
step:1295/1530 train_loss:3.2835 train_time:221688ms step_avg:172.52ms
step:1296/1530 train_loss:3.3758 train_time:221869ms step_avg:172.53ms
step:1297/1530 train_loss:3.2788 train_time:222048ms step_avg:172.53ms
step:1298/1530 train_loss:3.2690 train_time:222228ms step_avg:172.54ms
step:1299/1530 train_loss:3.3940 train_time:222406ms step_avg:172.54ms
step:1300/1530 train_loss:3.4011 train_time:222582ms step_avg:172.54ms
step:1301/1530 train_loss:3.4058 train_time:222757ms step_avg:172.55ms
step:1302/1530 train_loss:3.5795 train_time:222938ms step_avg:172.55ms
step:1303/1530 train_loss:3.3062 train_time:223121ms step_avg:172.56ms
step:1304/1530 train_loss:3.5104 train_time:223302ms step_avg:172.57ms
step:1305/1530 train_loss:3.2572 train_time:223478ms step_avg:172.57ms
step:1306/1530 train_loss:3.4526 train_time:223659ms step_avg:172.58ms
step:1307/1530 train_loss:3.4556 train_time:223834ms step_avg:172.58ms
step:1308/1530 train_loss:3.2874 train_time:224014ms step_avg:172.58ms
step:1309/1530 train_loss:3.3115 train_time:224195ms step_avg:172.59ms
step:1310/1530 train_loss:3.2917 train_time:224374ms step_avg:172.60ms
step:1311/1530 train_loss:3.2978 train_time:224552ms step_avg:172.60ms
step:1312/1530 train_loss:3.3766 train_time:224732ms step_avg:172.61ms
step:1313/1530 train_loss:3.3410 train_time:224909ms step_avg:172.61ms
step:1314/1530 train_loss:3.0441 train_time:225092ms step_avg:172.62ms
step:1315/1530 train_loss:3.2798 train_time:225269ms step_avg:172.62ms
step:1316/1530 train_loss:3.3970 train_time:225445ms step_avg:172.62ms
step:1317/1530 train_loss:3.4199 train_time:225622ms step_avg:172.63ms
step:1318/1530 train_loss:3.3023 train_time:225807ms step_avg:172.64ms
step:1319/1530 train_loss:3.4325 train_time:225987ms step_avg:172.64ms
step:1320/1530 train_loss:3.4669 train_time:226169ms step_avg:172.65ms
step:1321/1530 train_loss:3.3653 train_time:226349ms step_avg:172.65ms
step:1322/1530 train_loss:3.3228 train_time:226662ms step_avg:172.76ms
step:1323/1530 train_loss:3.3201 train_time:226851ms step_avg:172.77ms
step:1324/1530 train_loss:3.4379 train_time:227031ms step_avg:172.78ms
step:1325/1530 train_loss:3.4979 train_time:227216ms step_avg:172.79ms
step:1326/1530 train_loss:3.2159 train_time:227397ms step_avg:172.79ms
step:1327/1530 train_loss:3.1668 train_time:227574ms step_avg:172.80ms
step:1328/1530 train_loss:3.4974 train_time:227754ms step_avg:172.80ms
step:1329/1530 train_loss:3.3005 train_time:228091ms step_avg:172.93ms
step:1330/1530 train_loss:3.4309 train_time:228273ms step_avg:172.93ms
step:1331/1530 train_loss:3.3351 train_time:228449ms step_avg:172.94ms
step:1332/1530 train_loss:3.7429 train_time:228629ms step_avg:172.94ms
step:1333/1530 train_loss:3.4806 train_time:228809ms step_avg:172.95ms
step:1334/1530 train_loss:3.3732 train_time:228988ms step_avg:172.95ms
step:1335/1530 train_loss:3.2961 train_time:229166ms step_avg:172.96ms
step:1336/1530 train_loss:3.2984 train_time:229350ms step_avg:172.96ms
step:1337/1530 train_loss:3.5522 train_time:229529ms step_avg:172.97ms
step:1338/1530 train_loss:3.5257 train_time:229709ms step_avg:172.97ms
step:1339/1530 train_loss:3.3352 train_time:229887ms step_avg:172.98ms
step:1340/1530 train_loss:3.2893 train_time:230067ms step_avg:172.98ms
step:1341/1530 train_loss:3.5962 train_time:230244ms step_avg:172.99ms
step:1342/1530 train_loss:3.3574 train_time:230422ms step_avg:172.99ms
step:1343/1530 train_loss:3.3655 train_time:230600ms step_avg:172.99ms
step:1344/1530 train_loss:3.4192 train_time:230779ms step_avg:173.00ms
step:1345/1530 train_loss:3.3819 train_time:230961ms step_avg:173.00ms
step:1346/1530 train_loss:3.2972 train_time:231137ms step_avg:173.01ms
step:1347/1530 train_loss:3.2813 train_time:231315ms step_avg:173.01ms
step:1348/1530 train_loss:3.3509 train_time:231493ms step_avg:173.01ms
step:1349/1530 train_loss:3.2764 train_time:231671ms step_avg:173.02ms
step:1350/1530 train_loss:3.3917 train_time:231852ms step_avg:173.02ms
step:1351/1530 train_loss:3.2455 train_time:232028ms step_avg:173.03ms
step:1352/1530 train_loss:3.3091 train_time:232208ms step_avg:173.03ms
step:1353/1530 train_loss:3.4067 train_time:232387ms step_avg:173.04ms
step:1354/1530 train_loss:3.2623 train_time:232565ms step_avg:173.04ms
step:1355/1530 train_loss:3.1911 train_time:232741ms step_avg:173.04ms
step:1356/1530 train_loss:3.5105 train_time:232921ms step_avg:173.05ms
step:1357/1530 train_loss:3.4278 train_time:233102ms step_avg:173.05ms
step:1358/1530 train_loss:3.1875 train_time:233279ms step_avg:173.06ms
step:1359/1530 train_loss:3.4429 train_time:233459ms step_avg:173.06ms
step:1360/1530 train_loss:3.3503 train_time:233638ms step_avg:173.07ms
step:1361/1530 train_loss:3.1256 train_time:233823ms step_avg:173.07ms
step:1362/1530 train_loss:3.3919 train_time:234004ms step_avg:173.08ms
step:1363/1530 train_loss:3.2818 train_time:234191ms step_avg:173.09ms
step:1364/1530 train_loss:3.3015 train_time:234371ms step_avg:173.10ms
step:1365/1530 train_loss:3.3166 train_time:234550ms step_avg:173.10ms
step:1366/1530 train_loss:3.4274 train_time:234731ms step_avg:173.11ms
step:1367/1530 train_loss:3.4015 train_time:234909ms step_avg:173.11ms
step:1368/1530 train_loss:3.3533 train_time:235090ms step_avg:173.11ms
step:1369/1530 train_loss:3.2817 train_time:235278ms step_avg:173.13ms
step:1370/1530 train_loss:3.6106 train_time:235457ms step_avg:173.13ms
step:1371/1530 train_loss:3.3125 train_time:235638ms step_avg:173.14ms
step:1372/1530 train_loss:3.3752 train_time:235822ms step_avg:173.14ms
step:1373/1530 train_loss:3.3722 train_time:236000ms step_avg:173.15ms
step:1374/1530 train_loss:3.1546 train_time:236181ms step_avg:173.15ms
step:1375/1530 train_loss:3.5365 train_time:236361ms step_avg:173.16ms
step:1375/1530 val_loss:3.3138 train_time:236411ms step_avg:173.20ms
step:1376/1530 train_loss:3.3481 train_time:236541ms step_avg:173.16ms
step:1377/1530 train_loss:3.4796 train_time:236719ms step_avg:173.17ms
step:1378/1530 train_loss:3.4629 train_time:236895ms step_avg:173.17ms
step:1379/1530 train_loss:3.1125 train_time:237078ms step_avg:173.18ms
step:1380/1530 train_loss:3.3131 train_time:237258ms step_avg:173.18ms
step:1381/1530 train_loss:3.7010 train_time:237443ms step_avg:173.19ms
step:1382/1530 train_loss:3.2123 train_time:237621ms step_avg:173.19ms
step:1383/1530 train_loss:3.3953 train_time:237803ms step_avg:173.20ms
step:1384/1530 train_loss:3.4774 train_time:237986ms step_avg:173.21ms
step:1385/1530 train_loss:3.4107 train_time:238161ms step_avg:173.21ms
step:1386/1530 train_loss:3.3422 train_time:238341ms step_avg:173.21ms
step:1387/1530 train_loss:3.2014 train_time:238519ms step_avg:173.22ms
step:1388/1530 train_loss:3.3489 train_time:238696ms step_avg:173.22ms
step:1389/1530 train_loss:3.3215 train_time:238880ms step_avg:173.23ms
step:1390/1530 train_loss:3.5699 train_time:239057ms step_avg:173.23ms
step:1391/1530 train_loss:3.2951 train_time:239234ms step_avg:173.23ms
step:1392/1530 train_loss:3.2915 train_time:239412ms step_avg:173.24ms
step:1393/1530 train_loss:3.2428 train_time:239592ms step_avg:173.24ms
step:1394/1530 train_loss:3.4997 train_time:239770ms step_avg:173.24ms
step:1395/1530 train_loss:3.3929 train_time:239947ms step_avg:173.25ms
step:1396/1530 train_loss:3.4062 train_time:240124ms step_avg:173.25ms
step:1397/1530 train_loss:3.3118 train_time:240300ms step_avg:173.25ms
step:1398/1530 train_loss:3.2557 train_time:240477ms step_avg:173.25ms
step:1399/1530 train_loss:3.3194 train_time:240656ms step_avg:173.26ms
step:1400/1530 train_loss:3.3228 train_time:240838ms step_avg:173.26ms
step:1401/1530 train_loss:3.3525 train_time:241014ms step_avg:173.27ms
step:1402/1530 train_loss:3.3004 train_time:241192ms step_avg:173.27ms
step:1403/1530 train_loss:3.4969 train_time:241375ms step_avg:173.28ms
step:1404/1530 train_loss:3.2819 train_time:241553ms step_avg:173.28ms
step:1405/1530 train_loss:3.3174 train_time:241734ms step_avg:173.29ms
step:1406/1530 train_loss:3.3160 train_time:241916ms step_avg:173.29ms
step:1407/1530 train_loss:3.1807 train_time:242090ms step_avg:173.29ms
step:1408/1530 train_loss:3.3143 train_time:242270ms step_avg:173.30ms
step:1409/1530 train_loss:3.3057 train_time:242458ms step_avg:173.31ms
step:1410/1530 train_loss:3.2879 train_time:242634ms step_avg:173.31ms
step:1411/1530 train_loss:3.3697 train_time:242811ms step_avg:173.31ms
step:1412/1530 train_loss:3.3336 train_time:242989ms step_avg:173.32ms
step:1413/1530 train_loss:3.3676 train_time:243168ms step_avg:173.32ms
step:1414/1530 train_loss:3.3318 train_time:243348ms step_avg:173.32ms
step:1415/1530 train_loss:3.4094 train_time:243533ms step_avg:173.33ms
step:1416/1530 train_loss:3.2254 train_time:243722ms step_avg:173.34ms
step:1417/1530 train_loss:3.2834 train_time:243905ms step_avg:173.35ms
step:1418/1530 train_loss:3.3903 train_time:244086ms step_avg:173.36ms
step:1419/1530 train_loss:3.3534 train_time:244268ms step_avg:173.36ms
step:1420/1530 train_loss:3.3667 train_time:244449ms step_avg:173.37ms
step:1421/1530 train_loss:3.3746 train_time:244629ms step_avg:173.37ms
step:1422/1530 train_loss:3.3423 train_time:244806ms step_avg:173.38ms
step:1423/1530 train_loss:3.3182 train_time:244986ms step_avg:173.38ms
step:1424/1530 train_loss:3.3354 train_time:245171ms step_avg:173.39ms
step:1425/1530 train_loss:3.1933 train_time:245358ms step_avg:173.40ms
step:1426/1530 train_loss:3.3219 train_time:245536ms step_avg:173.40ms
step:1427/1530 train_loss:3.2840 train_time:245719ms step_avg:173.41ms
step:1428/1530 train_loss:3.3791 train_time:245897ms step_avg:173.41ms
step:1429/1530 train_loss:3.3580 train_time:246073ms step_avg:173.41ms
step:1430/1530 train_loss:3.2626 train_time:246254ms step_avg:173.42ms
step:1431/1530 train_loss:3.3262 train_time:246434ms step_avg:173.42ms
step:1432/1530 train_loss:3.3427 train_time:246615ms step_avg:173.43ms
step:1433/1530 train_loss:3.1336 train_time:246798ms step_avg:173.44ms
step:1434/1530 train_loss:3.2878 train_time:246983ms step_avg:173.44ms
step:1435/1530 train_loss:3.1238 train_time:247163ms step_avg:173.45ms
step:1436/1530 train_loss:3.2335 train_time:247342ms step_avg:173.45ms
step:1437/1530 train_loss:3.4074 train_time:247519ms step_avg:173.45ms
step:1438/1530 train_loss:3.3826 train_time:247694ms step_avg:173.46ms
step:1439/1530 train_loss:3.3201 train_time:247876ms step_avg:173.46ms
step:1440/1530 train_loss:3.1925 train_time:248052ms step_avg:173.46ms
step:1441/1530 train_loss:3.3405 train_time:248231ms step_avg:173.47ms
step:1442/1530 train_loss:3.3942 train_time:248416ms step_avg:173.47ms
step:1443/1530 train_loss:3.4920 train_time:248603ms step_avg:173.48ms
step:1444/1530 train_loss:3.4500 train_time:248780ms step_avg:173.49ms
step:1445/1530 train_loss:3.3385 train_time:248955ms step_avg:173.49ms
step:1446/1530 train_loss:3.1976 train_time:249135ms step_avg:173.49ms
step:1447/1530 train_loss:3.3014 train_time:249317ms step_avg:173.50ms
step:1448/1530 train_loss:3.3008 train_time:249493ms step_avg:173.50ms
step:1449/1530 train_loss:3.3941 train_time:249673ms step_avg:173.50ms
step:1450/1530 train_loss:3.3889 train_time:249852ms step_avg:173.51ms
step:1451/1530 train_loss:3.2069 train_time:250032ms step_avg:173.51ms
step:1452/1530 train_loss:3.3274 train_time:250211ms step_avg:173.52ms
step:1453/1530 train_loss:3.2604 train_time:250386ms step_avg:173.52ms
step:1454/1530 train_loss:3.2920 train_time:250565ms step_avg:173.52ms
step:1455/1530 train_loss:3.3313 train_time:250747ms step_avg:173.53ms
step:1456/1530 train_loss:3.2813 train_time:250924ms step_avg:173.53ms
step:1457/1530 train_loss:3.1568 train_time:251103ms step_avg:173.53ms
step:1458/1530 train_loss:3.4221 train_time:251283ms step_avg:173.54ms
step:1459/1530 train_loss:3.2718 train_time:251466ms step_avg:173.54ms
step:1460/1530 train_loss:3.3210 train_time:251645ms step_avg:173.55ms
step:1461/1530 train_loss:3.4298 train_time:251825ms step_avg:173.55ms
step:1462/1530 train_loss:3.2637 train_time:252002ms step_avg:173.56ms
step:1463/1530 train_loss:3.4677 train_time:252186ms step_avg:173.56ms
step:1464/1530 train_loss:3.3620 train_time:252365ms step_avg:173.57ms
step:1465/1530 train_loss:3.3559 train_time:252547ms step_avg:173.57ms
step:1466/1530 train_loss:3.2894 train_time:252723ms step_avg:173.57ms
step:1467/1530 train_loss:3.3963 train_time:252904ms step_avg:173.58ms
step:1468/1530 train_loss:3.2914 train_time:253082ms step_avg:173.58ms
step:1469/1530 train_loss:3.2757 train_time:253262ms step_avg:173.59ms
step:1470/1530 train_loss:3.3327 train_time:253445ms step_avg:173.59ms
step:1471/1530 train_loss:3.2643 train_time:253630ms step_avg:173.60ms
step:1472/1530 train_loss:3.2538 train_time:253814ms step_avg:173.61ms
step:1473/1530 train_loss:3.4426 train_time:253992ms step_avg:173.61ms
step:1474/1530 train_loss:3.3160 train_time:254176ms step_avg:173.62ms
step:1475/1530 train_loss:3.1524 train_time:254364ms step_avg:173.63ms
step:1476/1530 train_loss:3.2667 train_time:254542ms step_avg:173.63ms
step:1477/1530 train_loss:3.2416 train_time:254730ms step_avg:173.64ms
step:1478/1530 train_loss:3.3157 train_time:254915ms step_avg:173.65ms
step:1479/1530 train_loss:3.3987 train_time:255097ms step_avg:173.65ms
step:1480/1530 train_loss:3.2690 train_time:255276ms step_avg:173.66ms
step:1481/1530 train_loss:3.4508 train_time:255456ms step_avg:173.66ms
step:1482/1530 train_loss:3.3730 train_time:255643ms step_avg:173.67ms
step:1483/1530 train_loss:3.2806 train_time:255834ms step_avg:173.68ms
step:1484/1530 train_loss:3.2686 train_time:256021ms step_avg:173.69ms
step:1485/1530 train_loss:3.2815 train_time:256199ms step_avg:173.69ms
step:1486/1530 train_loss:3.2269 train_time:256386ms step_avg:173.70ms
step:1487/1530 train_loss:3.3415 train_time:256569ms step_avg:173.71ms
step:1488/1530 train_loss:3.2461 train_time:256753ms step_avg:173.72ms
step:1489/1530 train_loss:3.3201 train_time:256934ms step_avg:173.72ms
step:1490/1530 train_loss:3.2509 train_time:257113ms step_avg:173.73ms
step:1491/1530 train_loss:3.1609 train_time:257293ms step_avg:173.73ms
step:1492/1530 train_loss:3.2705 train_time:257472ms step_avg:173.73ms
step:1493/1530 train_loss:3.4355 train_time:257652ms step_avg:173.74ms
step:1494/1530 train_loss:3.2986 train_time:257832ms step_avg:173.74ms
step:1495/1530 train_loss:3.0372 train_time:258018ms step_avg:173.75ms
step:1496/1530 train_loss:3.3614 train_time:258199ms step_avg:173.75ms
step:1497/1530 train_loss:3.3152 train_time:258385ms step_avg:173.76ms
step:1498/1530 train_loss:3.3486 train_time:258569ms step_avg:173.77ms
step:1499/1530 train_loss:3.3114 train_time:258758ms step_avg:173.78ms
step:1500/1530 train_loss:3.2979 train_time:258948ms step_avg:173.79ms
step:1500/1530 val_loss:3.2819 train_time:259002ms step_avg:173.83ms
step:1501/1530 train_loss:3.0877 train_time:259137ms step_avg:173.80ms
step:1502/1530 train_loss:3.3632 train_time:259330ms step_avg:173.81ms
step:1503/1530 train_loss:3.2445 train_time:259509ms step_avg:173.82ms
step:1504/1530 train_loss:3.2482 train_time:259694ms step_avg:173.82ms
step:1505/1530 train_loss:3.2111 train_time:259873ms step_avg:173.83ms
step:1506/1530 train_loss:3.2825 train_time:260057ms step_avg:173.83ms
step:1507/1530 train_loss:3.1835 train_time:260251ms step_avg:173.85ms
step:1508/1530 train_loss:3.4875 train_time:260434ms step_avg:173.85ms
step:1509/1530 train_loss:3.2821 train_time:260611ms step_avg:173.86ms
step:1510/1530 train_loss:3.2722 train_time:260792ms step_avg:173.86ms
step:1511/1530 train_loss:3.4154 train_time:261115ms step_avg:173.96ms
step:1512/1530 train_loss:3.4225 train_time:261303ms step_avg:173.97ms
step:1513/1530 train_loss:3.2701 train_time:261487ms step_avg:173.98ms
step:1514/1530 train_loss:3.0884 train_time:261670ms step_avg:173.98ms
step:1515/1530 train_loss:3.2435 train_time:261851ms step_avg:173.99ms
step:1516/1530 train_loss:3.2568 train_time:262036ms step_avg:173.99ms
step:1517/1530 train_loss:3.3021 train_time:262219ms step_avg:174.00ms
step:1518/1530 train_loss:3.2100 train_time:262401ms step_avg:174.01ms
step:1519/1530 train_loss:3.5054 train_time:262734ms step_avg:174.11ms
step:1520/1530 train_loss:3.1283 train_time:262914ms step_avg:174.12ms
step:1521/1530 train_loss:3.2098 train_time:263092ms step_avg:174.12ms
step:1522/1530 train_loss:3.3564 train_time:263277ms step_avg:174.13ms
step:1523/1530 train_loss:3.2343 train_time:263455ms step_avg:174.13ms
step:1524/1530 train_loss:3.3520 train_time:263637ms step_avg:174.13ms
step:1525/1530 train_loss:3.3402 train_time:263824ms step_avg:174.14ms
step:1526/1530 train_loss:3.2789 train_time:264013ms step_avg:174.15ms
step:1527/1530 train_loss:3.2906 train_time:264195ms step_avg:174.16ms
step:1528/1530 train_loss:3.4102 train_time:264375ms step_avg:174.16ms
step:1529/1530 train_loss:3.4100 train_time:264552ms step_avg:174.16ms
step:1530/1530 train_loss:3.2373 train_time:264731ms step_avg:174.17ms
step:1530/1530 val_loss:3.2794 train_time:264785ms step_avg:174.20ms