# 使用 Keras 中的再訓練 VGG16 進行圖像分類 讓我們使用 COCO 圖像數據集來再訓練模型以微調分類任務。我們將刪除 Keras 模型中的最后一層，并添加我們自己的完全連接層，其中`softmax`激活 8 個類。我們還將通過將前 15 層的`trainable`屬性設置為`False`來演示凍結前幾層。 1. 首先導入 VGG16 模型而不使用頂層變量，方法是將`include_top`設置為`False`： ```py # load the vgg model from keras.applications import VGG16 base_model=VGG16(weights='imagenet',include_top=False, input_shape=(224,224,3)) ``` 我們還在上面的代碼中指定了`input_shape`，否則 Keras 會在以后拋出異常。 1. 現在我們構建分類器模型以置于導入的 VGG 模型之上： ```py top_model = Sequential() top_model.add(Flatten(input_shape=base_model.output_shape[1:])) top_model.add(Dense(256, activation='relu')) top_model.add(Dropout(0.5)) top_model.add(Dense(coco.n_classes, activation='softmax')) ``` 1. 接下來，在 VGG 基礎之上添加模型： ```py model=Model(inputs=base_model.input, outputs=top_model(base_model.output)) ``` 1. 凍結前 15 層： ```py for layer in model.layers[:15]: layer.trainable = False ``` 1. 我們隨機挑選了 15 層凍結，你可能想要玩這個數字。讓我們編譯模型并打印模型摘要： ```py model.compile(loss='categorical_crossentropy', optimizer=optimizers.SGD(lr=1e-4, momentum=0.9), metrics=['accuracy']) model.summary() ``` ```py _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= input_1 (InputLayer) (None, 224, 224, 3) 0 _________________________________________________________________ block1_conv1 (Conv2D) (None, 224, 224, 64) 1792 _________________________________________________________________ block1_conv2 (Conv2D) (None, 224, 224, 64) 36928 _________________________________________________________________ block1_pool (MaxPooling2D) (None, 112, 112, 64) 0 _________________________________________________________________ block2_conv1 (Conv2D) (None, 112, 112, 128) 73856 _________________________________________________________________ block2_conv2 (Conv2D) (None, 112, 112, 128) 147584 _________________________________________________________________ block2_pool (MaxPooling2D) (None, 56, 56, 128) 0 _________________________________________________________________ block3_conv1 (Conv2D) (None, 56, 56, 256) 295168 _________________________________________________________________ block3_conv2 (Conv2D) (None, 56, 56, 256) 590080 _________________________________________________________________ block3_conv3 (Conv2D) (None, 56, 56, 256) 590080 _________________________________________________________________ block3_pool (MaxPooling2D) (None, 28, 28, 256) 0 _________________________________________________________________ block4_conv1 (Conv2D) (None, 28, 28, 512) 1180160 _________________________________________________________________ block4_conv2 (Conv2D) (None, 28, 28, 512) 2359808 _________________________________________________________________ block4_conv3 (Conv2D) (None, 28, 28, 512) 2359808 _________________________________________________________________ block4_pool (MaxPooling2D) (None, 14, 14, 512) 0 _________________________________________________________________ block5_conv1 (Conv2D) (None, 14, 14, 512) 2359808 _________________________________________________________________ block5_conv2 (Conv2D) (None, 14, 14, 512) 2359808 _________________________________________________________________ block5_conv3 (Conv2D) (None, 14, 14, 512) 2359808 _________________________________________________________________ block5_pool (MaxPooling2D) (None, 7, 7, 512) 0 _________________________________________________________________ sequential_1 (Sequential) (None, 8) 6424840 ================================================================= Total params: 21,139,528 Trainable params: 13,504,264 Non-trainable params: 7,635,264 ``` 我們看到近 40%的參數是凍結的和不可訓練的。 1. 接下來，訓練 Keras 模型 20 個周期，批量大小為 32： ```py from keras.utils import np_utils batch_size=32 n_epochs=20 total_images = len(x_train_files) n_batches = total_images // batch_size for epoch in range(n_epochs): print('Starting epoch ',epoch) coco.reset_index_in_epoch() for batch in range(n_batches): try: x_batch, y_batch = coco.next_batch(batch_size=batch_size) images=np.array([coco.preprocess_image(x) for x in x_batch]) y_onehot = np_utils.to_categorical(y_batch, num_classes=coco.n_classes) model.fit(x=images,y=y_onehot,verbose=0) except Exception as ex: print('error in epoch {} batch {}'.format(epoch,batch)) print(ex) ``` 1. 讓我們使用新再訓練的模型對圖像進行分類： ```py probs = model.predict(images_test) ``` 以下是分類結果：  ```py Probability 100.00% of [zebra] Probability 0.00% of [dog] Probability 0.00% of [horse] Probability 0.00% of [giraffe] Probability 0.00% of [bear] ``` ---  ```py Probability 96.11% of [horse] Probability 1.85% of [cat] Probability 0.77% of [bird] Probability 0.43% of [giraffe] Probability 0.40% of [sheep] ``` ---  ```py Probability 99.75% of [dog] Probability 0.22% of [cat] Probability 0.03% of [horse] Probability 0.00% of [bear] Probability 0.00% of [zebra] ``` ---  ```py Probability 99.88% of [bird] Probability 0.11% of [horse] Probability 0.00% of [giraffe] Probability 0.00% of [bear] Probability 0.00% of [cat] ``` ---  ```py Probability 65.28% of [bear] Probability 27.09% of [sheep] Probability 4.34% of [bird] Probability 1.71% of [giraffe] Probability 0.63% of [dog] ``` ---  ```py Probability 100.00% of [bear] Probability 0.00% of [sheep] Probability 0.00% of [dog] Probability 0.00% of [cat] Probability 0.00% of [giraffe] ``` ---  ```py Probability 100.00% of [giraffe] Probability 0.00% of [bird] Probability 0.00% of [bear] Probability 0.00% of [sheep] Probability 0.00% of [zebra] ``` ---  ```py Probability 81.05% of [cat] Probability 15.68% of [dog] Probability 1.64% of [bird] Probability 0.90% of [horse] Probability 0.43% of [bear] ``` 除了最后的嘈雜圖像外，所有類別都已正確識別。通過適當的超參數調整，也可以進行改進。 到目前為止，您已經看到了使用預訓練模型進行分類并對預訓練模型進行微調的示例。接下來，我們將使用 Inception v3 模型顯示分類示例。