Merge pull request #5 from restbai/workshop

Notebook for the workshop - final version of the notebook

workshop/places365_train.ipynb

@@ -17,7 +17,9 @@
       },
       "outputs": [],
       "source": [
-        "# !gdown 1Hkk2HNvnh2cZqIcOGuxpxUkDSDh-QW86\n",
+        "# 1200:\n",
+        "#!gdown 1Hkk2HNvnh2cZqIcOGuxpxUkDSDh-QW86\n",
+        "# 300:\n",
         "!gdown 1y-LdQ_4dbOip6sBgZ-Ub1FI6Hh5kl3h1"
       ]
     },
@@ -189,33 +191,24 @@
       },
       "outputs": [],
       "source": [
-        "class Places365Model(tf.keras.Model):\n",
-        "  def __init__(self):\n",
-        "    super().__init__()\n",
-        "    self.base_model = tf.keras.applications.MobileNetV3Small(\n",
+        "inputs = tf.keras.Input(shape=(224, 224, 3))\n",
+        "x = tf.keras.applications.MobileNetV3Small(\n",
         "      input_shape=(224, 224, 3),\n",
         "      include_top=False,\n",
         "      weights=None,\n",
         "      classes=3,\n",
         "      pooling=\"avg\",\n",
         "      minimalistic=True\n",
-        "    )\n",
-        "    self.fc = tf.keras.layers.Dense(3, activation='softmax')\n",
-        "\n",
-        "  def call(self, x):\n",
-        "    x = self.base_model(x)\n",
-        "    return self.fc(x)\n",
-        "\n",
-        "\n",
-        "model = Places365Model()\n",
+        "    )(inputs)\n",
+        "outputs = tf.keras.layers.Dense(3, activation=\"softmax\")(x)\n",
+        "model = tf.keras.Model(inputs=inputs, outputs=outputs)\n",
         "\n",
         "# updating momentum of the BatchNorm layers\n",
-        "for layer in model.base_model.layers:\n",
+        "for layer in model.layers:\n",
         "  if isinstance(layer, tf.keras.layers.BatchNormalization):\n",
         "    layer.momentum = 0.5\n",
         "\n",
-        "optimizer = tf.keras.optimizers.Adam(learning_rate=0.005)\n",
-        "\n",
+        "optimizer = tf.keras.optimizers.Adam(learning_rate=0.001)\n",
         "model.compile(optimizer=optimizer, loss='sparse_categorical_crossentropy', metrics='accuracy')"
       ]
     },
@@ -238,7 +231,7 @@
       "source": [
         "history = model.fit(\n",
         "    train_ds,\n",
-        "    epochs=32)"
+        "    epochs=20)"
       ]
     },
     {
@@ -399,17 +392,26 @@
         "  img, label = gt\n",
         "  predicted_class = np.argmax(result)\n",
         "  if predicted_class != label:\n",
-        "    images_to_plot.append((img, label, predicted_class))\n",
-        "\n",
+        "    images_to_plot.append((img, label, predicted_class))"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "source": [
         "fig = plt.figure(figsize=(128., 128.))\n",
         "grid = ImageGrid(fig, 111, nrows_ncols=(math.ceil(len(images_to_plot) / 4), 4), axes_pad=0.6,)\n",
         "\n",
         "for ax, im in zip(grid, images_to_plot):\n",
-        "    ax.set_title(f\"True: {class_names[im[1]]}, predicted: {class_names[im[2]]}\", fontdict=None, loc='center', color = \"k\", fontsize=15)\n",
+        "    ax.set_title(f\"True: {class_names[im[1]]}, predicted: {class_names[im[2]]}\", fontdict=None, loc='center', color = \"k\", fontsize=70)\n",
         "    ax.imshow(im[0] / 255)\n",
         "\n",
         "plt.show()"
-      ]
+      ],
+      "metadata": {
+        "id": "J8plRcodbaXT"
+      },
+      "execution_count": null,
+      "outputs": []
     }
   ],
   "metadata": {