arthur/handson-ml3
1
0
Fork 0
mirror of https://github.com/ArthurDanjou/handson-ml3.git synced 2026-01-29 19:20:28 +01:00
Code Issues Packages Projects Releases Wiki Activity

SGD now defaults to lr=0.01 so use 1e-3 explicitely

Browse Source
This commit is contained in:
Aurélien Geron
2019-06-10 10:48:00 +08:00
parent 1973371b19
commit 3db31444cd
1 changed files with 21 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

32
11_training_deep_neural_networks.ipynb
View File

@@ -511,7 +511,8 @@
"metadata": {}, "metadata": {},
"outputs": [], "outputs": [],
"source": [ "source": [
"model.compile(loss=\"sparse_categorical_crossentropy\", optimizer=\"sgd\",\n", "model.compile(loss=\"sparse_categorical_crossentropy\",\n",
" optimizer=keras.optimizers.SGD(lr=1e-3),\n",
" metrics=[\"accuracy\"])" " metrics=[\"accuracy\"])"
] ]
}, },
@@ -582,7 +583,8 @@
"metadata": {}, "metadata": {},
"outputs": [], "outputs": [],
"source": [ "source": [
"model.compile(loss=\"sparse_categorical_crossentropy\", optimizer=\"sgd\",\n", "model.compile(loss=\"sparse_categorical_crossentropy\",\n",
" optimizer=keras.optimizers.SGD(lr=1e-3),\n",
" metrics=[\"accuracy\"])" " metrics=[\"accuracy\"])"
] ]
}, },
@@ -661,7 +663,8 @@
"metadata": {}, "metadata": {},
"outputs": [], "outputs": [],
"source": [ "source": [
"model.compile(loss=\"sparse_categorical_crossentropy\", optimizer=\"sgd\",\n", "model.compile(loss=\"sparse_categorical_crossentropy\",\n",
" optimizer=keras.optimizers.SGD(lr=1e-3),\n",
" metrics=[\"accuracy\"])" " metrics=[\"accuracy\"])"
] ]
}, },
@@ -707,7 +710,8 @@
"metadata": {}, "metadata": {},
"outputs": [], "outputs": [],
"source": [ "source": [
"model.compile(loss=\"sparse_categorical_crossentropy\", optimizer=\"sgd\",\n", "model.compile(loss=\"sparse_categorical_crossentropy\",\n",
" optimizer=keras.optimizers.SGD(lr=1e-3),\n",
" metrics=[\"accuracy\"])" " metrics=[\"accuracy\"])"
] ]
}, },
@@ -866,8 +870,9 @@
"metadata": {}, "metadata": {},
"outputs": [], "outputs": [],
"source": [ "source": [
"model_A.compile(loss=\"sparse_categorical_crossentropy\", optimizer=\"sgd\",\n", "model_A.compile(loss=\"sparse_categorical_crossentropy\",\n",
" metrics=[\"accuracy\"])" " optimizer=keras.optimizers.SGD(lr=1e-3),\n",
" metrics=[\"accuracy\"])"
] ]
}, },
{ {
@@ -908,8 +913,9 @@
"metadata": {}, "metadata": {},
"outputs": [], "outputs": [],
"source": [ "source": [
"model_B.compile(loss=\"binary_crossentropy\", optimizer=\"sgd\",\n", "model_B.compile(loss=\"binary_crossentropy\",\n",
" metrics=[\"accuracy\"])" " optimizer=keras.optimizers.SGD(lr=1e-3),\n",
" metrics=[\"accuracy\"])"
] ]
}, },
{ {
@@ -961,7 +967,8 @@
"for layer in model_B_on_A.layers[:-1]:\n", "for layer in model_B_on_A.layers[:-1]:\n",
" layer.trainable = False\n", " layer.trainable = False\n",
"\n", "\n",
"model_B_on_A.compile(loss=\"binary_crossentropy\", optimizer=\"sgd\",\n", "model_B_on_A.compile(loss=\"binary_crossentropy\",\n",
" optimizer=keras.optimizers.SGD(lr=1e-3),\n",
" metrics=[\"accuracy\"])" " metrics=[\"accuracy\"])"
] ]
}, },
@@ -977,7 +984,8 @@
"for layer in model_B_on_A.layers[:-1]:\n", "for layer in model_B_on_A.layers[:-1]:\n",
" layer.trainable = True\n", " layer.trainable = True\n",
"\n", "\n",
"model_B_on_A.compile(loss=\"binary_crossentropy\", optimizer=\"sgd\",\n", "model_B_on_A.compile(loss=\"binary_crossentropy\",\n",
" optimizer=keras.optimizers.SGD(lr=1e-3),\n",
" metrics=[\"accuracy\"])\n", " metrics=[\"accuracy\"])\n",
"history = model_B_on_A.fit(X_train_B, y_train_B, epochs=16,\n", "history = model_B_on_A.fit(X_train_B, y_train_B, epochs=16,\n",
" validation_data=(X_valid_B, y_valid_B))" " validation_data=(X_valid_B, y_valid_B))"
@@ -1638,7 +1646,9 @@
" keras.layers.Dense(100, activation=\"selu\", kernel_initializer=\"lecun_normal\"),\n", " keras.layers.Dense(100, activation=\"selu\", kernel_initializer=\"lecun_normal\"),\n",
" keras.layers.Dense(10, activation=\"softmax\")\n", " keras.layers.Dense(10, activation=\"softmax\")\n",
"])\n", "])\n",
"model.compile(loss=\"sparse_categorical_crossentropy\", optimizer=\"sgd\", metrics=[\"accuracy\"])" "model.compile(loss=\"sparse_categorical_crossentropy\",\n",
" optimizer=keras.optimizers.SGD(lr=1e-3),\n",
" metrics=[\"accuracy\"])"
] ]
}, },
{ {