"
- ]
- },
- "execution_count": 65,
- "metadata": {},
- "output_type": "execute_result"
- }
- ],
+ "outputs": [],
"source": [
"# Include the epoch in the file name (uses `str.format`)\n",
"checkpoint_path = \"training_2/cp-{epoch:04d}.weights.h5\"\n",
@@ -655,57 +424,22 @@
},
{
"cell_type": "code",
- "execution_count": 66,
+ "execution_count": null,
"metadata": {
"id": "p64q3-V4sXt0"
},
- "outputs": [
- {
- "data": {
- "text/plain": [
- "['cp-0035.weights.h5',\n",
- " 'cp-0000.weights.h5',\n",
- " 'cp-0030.weights.h5',\n",
- " 'cp-0025.weights.h5',\n",
- " 'cp-0050.weights.h5',\n",
- " 'cp-0015.weights.h5',\n",
- " 'cp-0005.weights.h5',\n",
- " 'cp-0040.weights.h5',\n",
- " 'cp-0010.weights.h5',\n",
- " 'cp-0020.weights.h5',\n",
- " 'cp-0045.weights.h5']"
- ]
- },
- "execution_count": 66,
- "metadata": {},
- "output_type": "execute_result"
- }
- ],
+ "outputs": [],
"source": [
"os.listdir(checkpoint_dir)"
]
},
{
"cell_type": "code",
- "execution_count": 67,
+ "execution_count": null,
"metadata": {
"id": "1AN_fnuyR41H"
},
- "outputs": [
- {
- "data": {
- "application/vnd.google.colaboratory.intrinsic+json": {
- "type": "string"
- },
- "text/plain": [
- "'training_2/cp-0050.weights.h5'"
- ]
- },
- "execution_count": 67,
- "metadata": {},
- "output_type": "execute_result"
- }
- ],
+ "outputs": [],
"source": [
"def load_latest_checkpoint(checkpoint_dir):\n",
" latest = max(os.listdir(checkpoint_dir), key=lambda f: int(f.split('-')[1].split('.')[0]))\n",
@@ -728,20 +462,11 @@
},
{
"cell_type": "code",
- "execution_count": 68,
+ "execution_count": null,
"metadata": {
"id": "3M04jyK-H3QK"
},
- "outputs": [
- {
- "name": "stdout",
- "output_type": "stream",
- "text": [
- "32/32 - 1s - 34ms/step - loss: 0.4968 - sparse_categorical_accuracy: 0.8730\n",
- "Restored model, accuracy: 87.30%\n"
- ]
- }
- ],
+ "outputs": [],
"source": [
"# Create a new model instance\n",
"model = create_model()\n",
@@ -789,20 +514,11 @@
},
{
"cell_type": "code",
- "execution_count": 69,
+ "execution_count": null,
"metadata": {
"id": "R7W5plyZ-u9X"
},
- "outputs": [
- {
- "name": "stdout",
- "output_type": "stream",
- "text": [
- "32/32 - 1s - 23ms/step - loss: 0.4968 - sparse_categorical_accuracy: 0.8730\n",
- "Restored model, accuracy: 87.30%\n"
- ]
- }
- ],
+ "outputs": [],
"source": [
"# Save the weights\n",
"model.save_weights('./checkpoints/my_checkpoint.weights.h5')\n",
@@ -869,28 +585,11 @@
},
{
"cell_type": "code",
- "execution_count": 70,
+ "execution_count": null,
"metadata": {
"id": "3f55mAXwukUX"
},
- "outputs": [
- {
- "name": "stdout",
- "output_type": "stream",
- "text": [
- "Epoch 1/5\n",
- "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 42ms/step - loss: 1.5987 - sparse_categorical_accuracy: 0.5177\n",
- "Epoch 2/5\n",
- "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 3ms/step - loss: 0.4480 - sparse_categorical_accuracy: 0.8843\n",
- "Epoch 3/5\n",
- "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 3ms/step - loss: 0.2776 - sparse_categorical_accuracy: 0.9279\n",
- "Epoch 4/5\n",
- "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 3ms/step - loss: 0.2168 - sparse_categorical_accuracy: 0.9447\n",
- "Epoch 5/5\n",
- "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 3ms/step - loss: 0.1379 - sparse_categorical_accuracy: 0.9763\n"
- ]
- }
- ],
+ "outputs": [],
"source": [
"# Create and train a new model instance.\n",
"model = create_model()\n",
@@ -911,106 +610,11 @@
},
{
"cell_type": "code",
- "execution_count": 71,
+ "execution_count": null,
"metadata": {
"id": "HyfUMOZwux_-"
},
- "outputs": [
- {
- "data": {
- "text/html": [
- "Model: \"sequential_23\"\n",
- "
\n"
- ],
- "text/plain": [
- "\u001b[1mModel: \"sequential_23\"\u001b[0m\n"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- },
- {
- "data": {
- "text/html": [
- "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┓\n",
- "┃ Layer (type) ┃ Output Shape ┃ Param # ┃\n",
- "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━┩\n",
- "│ dense_46 (Dense) │ (None, 512) │ 401,920 │\n",
- "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
- "│ dropout_23 (Dropout) │ (None, 512) │ 0 │\n",
- "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
- "│ dense_47 (Dense) │ (None, 10) │ 5,130 │\n",
- "└──────────────────────────────────────┴─────────────────────────────┴─────────────────┘\n",
- "
\n"
- ],
- "text/plain": [
- "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┓\n",
- "┃\u001b[1m \u001b[0m\u001b[1mLayer (type) \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
- "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━┩\n",
- "│ dense_46 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m) │ \u001b[38;5;34m401,920\u001b[0m │\n",
- "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
- "│ dropout_23 (\u001b[38;5;33mDropout\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m) │ \u001b[38;5;34m0\u001b[0m │\n",
- "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
- "│ dense_47 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m10\u001b[0m) │ \u001b[38;5;34m5,130\u001b[0m │\n",
- "└──────────────────────────────────────┴─────────────────────────────┴─────────────────┘\n"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- },
- {
- "data": {
- "text/html": [
- " Total params: 1,221,152 (4.66 MB)\n",
- "
\n"
- ],
- "text/plain": [
- "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m1,221,152\u001b[0m (4.66 MB)\n"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- },
- {
- "data": {
- "text/html": [
- " Trainable params: 407,050 (1.55 MB)\n",
- "
\n"
- ],
- "text/plain": [
- "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m407,050\u001b[0m (1.55 MB)\n"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- },
- {
- "data": {
- "text/html": [
- " Non-trainable params: 0 (0.00 B)\n",
- "
\n"
- ],
- "text/plain": [
- "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- },
- {
- "data": {
- "text/html": [
- " Optimizer params: 814,102 (3.11 MB)\n",
- "
\n"
- ],
- "text/plain": [
- "\u001b[1m Optimizer params: \u001b[0m\u001b[38;5;34m814,102\u001b[0m (3.11 MB)\n"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- }
- ],
+ "outputs": [],
"source": [
"new_model = tf.keras.models.load_model('my_model.keras')\n",
"\n",
@@ -1029,22 +633,11 @@
},
{
"cell_type": "code",
- "execution_count": 72,
+ "execution_count": null,
"metadata": {
"id": "8BT4mHNIvMdW"
},
- "outputs": [
- {
- "name": "stdout",
- "output_type": "stream",
- "text": [
- "32/32 - 1s - 33ms/step - loss: 0.4288 - sparse_categorical_accuracy: 0.8620\n",
- "Restored model, accuracy: 86.20%\n",
- "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 3ms/step\n",
- "(1000, 10)\n"
- ]
- }
- ],
+ "outputs": [],
"source": [
"# Evaluate the restored model\n",
"loss, acc = new_model.evaluate(test_images, test_labels, verbose=2)\n",
@@ -1073,42 +666,11 @@
},
{
"cell_type": "code",
- "execution_count": 76,
+ "execution_count": null,
"metadata": {
"id": "sI1YvCDFzpl3"
},
- "outputs": [
- {
- "name": "stdout",
- "output_type": "stream",
- "text": [
- "Epoch 1/5\n"
- ]
- },
- {
- "name": "stderr",
- "output_type": "stream",
- "text": [
- "/usr/local/lib/python3.10/dist-packages/keras/src/layers/core/dense.py:87: UserWarning: Do not pass an `input_shape`/`input_dim` argument to a layer. When using Sequential models, prefer using an `Input(shape)` object as the first layer in the model instead.\n",
- " super().__init__(activity_regularizer=activity_regularizer, **kwargs)\n"
- ]
- },
- {
- "name": "stdout",
- "output_type": "stream",
- "text": [
- "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 30ms/step - loss: 1.5705 - sparse_categorical_accuracy: 0.5231\n",
- "Epoch 2/5\n",
- "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - loss: 0.4329 - sparse_categorical_accuracy: 0.8752 \n",
- "Epoch 3/5\n",
- "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - loss: 0.2856 - sparse_categorical_accuracy: 0.9221 \n",
- "Epoch 4/5\n",
- "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - loss: 0.1993 - sparse_categorical_accuracy: 0.9594 \n",
- "Epoch 5/5\n",
- "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - loss: 0.1284 - sparse_categorical_accuracy: 0.9760 \n"
- ]
- }
- ],
+ "outputs": [],
"source": [
"# Create and train a new model instance.\n",
"model = create_model()\n",
@@ -1130,20 +692,11 @@
},
{
"cell_type": "code",
- "execution_count": 77,
+ "execution_count": null,
"metadata": {
"id": "sq8fPglI1RWA"
},
- "outputs": [
- {
- "name": "stdout",
- "output_type": "stream",
- "text": [
- "my_model\n",
- "assets\tfingerprint.pb\tsaved_model.pb\tvariables\n"
- ]
- }
- ],
+ "outputs": [],
"source": [
"# my_model directory\n",
"!ls saved_model\n",
@@ -1163,22 +716,11 @@
},
{
"cell_type": "code",
- "execution_count": 78,
+ "execution_count": null,
"metadata": {
"id": "0YofwHdN0pxa"
},
- "outputs": [
- {
- "data": {
- "text/plain": [
- "._UserObject at 0x7c84e32913c0>"
- ]
- },
- "execution_count": 78,
- "metadata": {},
- "output_type": "execute_result"
- }
- ],
+ "outputs": [],
"source": [
"saved_model = tf.saved_model.load('saved_model/my_model')\n",
"saved_model"
@@ -1197,49 +739,11 @@
},
{
"cell_type": "code",
- "execution_count": 79,
+ "execution_count": null,
"metadata": {
"id": "m2dkmJVCGUia"
},
- "outputs": [
- {
- "name": "stdout",
- "output_type": "stream",
- "text": [
- "Epoch 1/5\n"
- ]
- },
- {
- "name": "stderr",
- "output_type": "stream",
- "text": [
- "/usr/local/lib/python3.10/dist-packages/keras/src/layers/core/dense.py:87: UserWarning: Do not pass an `input_shape`/`input_dim` argument to a layer. When using Sequential models, prefer using an `Input(shape)` object as the first layer in the model instead.\n",
- " super().__init__(activity_regularizer=activity_regularizer, **kwargs)\n"
- ]
- },
- {
- "name": "stdout",
- "output_type": "stream",
- "text": [
- "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 31ms/step - loss: 1.6114 - sparse_categorical_accuracy: 0.5003\n",
- "Epoch 2/5\n",
- "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 2ms/step - loss: 0.4265 - sparse_categorical_accuracy: 0.8682\n",
- "Epoch 3/5\n",
- "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 1ms/step - loss: 0.3224 - sparse_categorical_accuracy: 0.9086 \n",
- "Epoch 4/5\n",
- "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - loss: 0.2046 - sparse_categorical_accuracy: 0.9593 \n",
- "Epoch 5/5\n",
- "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - loss: 0.1646 - sparse_categorical_accuracy: 0.9715\n"
- ]
- },
- {
- "name": "stderr",
- "output_type": "stream",
- "text": [
- "WARNING:absl:You are saving your model as an HDF5 file via `model.save()` or `keras.saving.save_model(model)`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')` or `keras.saving.save_model(model, 'my_model.keras')`. \n"
- ]
- }
- ],
+ "outputs": [],
"source": [
"# Create and train a new model instance.\n",
"model = create_model()\n",
@@ -1261,113 +765,11 @@
},
{
"cell_type": "code",
- "execution_count": 80,
+ "execution_count": null,
"metadata": {
"id": "5NDMO_7kS6Do"
},
- "outputs": [
- {
- "name": "stderr",
- "output_type": "stream",
- "text": [
- "WARNING:absl:Compiled the loaded model, but the compiled metrics have yet to be built. `model.compile_metrics` will be empty until you train or evaluate the model.\n"
- ]
- },
- {
- "data": {
- "text/html": [
- "Model: \"sequential_28\"\n",
- "
\n"
- ],
- "text/plain": [
- "\u001b[1mModel: \"sequential_28\"\u001b[0m\n"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- },
- {
- "data": {
- "text/html": [
- "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┓\n",
- "┃ Layer (type) ┃ Output Shape ┃ Param # ┃\n",
- "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━┩\n",
- "│ dense_56 (Dense) │ (None, 512) │ 401,920 │\n",
- "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
- "│ dropout_28 (Dropout) │ (None, 512) │ 0 │\n",
- "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
- "│ dense_57 (Dense) │ (None, 10) │ 5,130 │\n",
- "└──────────────────────────────────────┴─────────────────────────────┴─────────────────┘\n",
- "
\n"
- ],
- "text/plain": [
- "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┓\n",
- "┃\u001b[1m \u001b[0m\u001b[1mLayer (type) \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
- "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━┩\n",
- "│ dense_56 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m) │ \u001b[38;5;34m401,920\u001b[0m │\n",
- "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
- "│ dropout_28 (\u001b[38;5;33mDropout\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m) │ \u001b[38;5;34m0\u001b[0m │\n",
- "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
- "│ dense_57 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m10\u001b[0m) │ \u001b[38;5;34m5,130\u001b[0m │\n",
- "└──────────────────────────────────────┴─────────────────────────────┴─────────────────┘\n"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- },
- {
- "data": {
- "text/html": [
- " Total params: 407,052 (1.55 MB)\n",
- "
\n"
- ],
- "text/plain": [
- "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m407,052\u001b[0m (1.55 MB)\n"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- },
- {
- "data": {
- "text/html": [
- " Trainable params: 407,050 (1.55 MB)\n",
- "
\n"
- ],
- "text/plain": [
- "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m407,050\u001b[0m (1.55 MB)\n"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- },
- {
- "data": {
- "text/html": [
- " Non-trainable params: 0 (0.00 B)\n",
- "
\n"
- ],
- "text/plain": [
- "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- },
- {
- "data": {
- "text/html": [
- " Optimizer params: 2 (12.00 B)\n",
- "
\n"
- ],
- "text/plain": [
- "\u001b[1m Optimizer params: \u001b[0m\u001b[38;5;34m2\u001b[0m (12.00 B)\n"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- }
- ],
+ "outputs": [],
"source": [
"# Recreate the exact same model, including its weights and the optimizer\n",
"new_model = tf.keras.models.load_model('my_model.h5')\n",
@@ -1387,20 +789,11 @@
},
{
"cell_type": "code",
- "execution_count": 81,
+ "execution_count": null,
"metadata": {
"id": "jwEaj9DnTCVA"
},
- "outputs": [
- {
- "name": "stdout",
- "output_type": "stream",
- "text": [
- "32/32 - 1s - 25ms/step - loss: 0.4343 - sparse_categorical_accuracy: 0.8610\n",
- "Restored model, accuracy: 86.10%\n"
- ]
- }
- ],
+ "outputs": [],
"source": [
"loss, acc = new_model.evaluate(test_images, test_labels, verbose=2)\n",
"print('Restored model, accuracy: {:5.2f}%'.format(100 * acc))"