Add keras mnist demo

demo/keras/TUTORIAL_CN.md

@@ -0,0 +1,90 @@
+# 如何在Keras中使用VisualDL
+
+下面我们演示一下如何在Keras中使用VisualDL，从而可以把Keras的训练过程可视化出来。我们将以Keras用卷积神经网络(CNN, Convolutional Neural Network)来训练
+[MNIST](http://yann.lecun.com/exdb/mnist/) 数据集作为例子。
+
+
+程序的主体来自Keras的官方GitHub [Example](https://github.com/keras-team/keras/blob/master/examples/mnist_cnn.py)。
+
+我们只需要在代码里面创建 VisualDL 的数据采集 loggers
+
+```python
+# create VisualDL logger
+logdir = "/workspace"
+logger = LogWriter(logdir, sync_cycle=100)
+
+# mark the components with 'train' label.
+with logger.mode("train"):
+    # create a scalar component called 'scalars/'
+    scalar_keras_train_loss = logger.scalar(
+        "scalars/scalar_keras_train_loss")
+    image0 = logger.image("images/image0", 1)
+    image1 = logger.image("images/image1", 1)
+    histogram0 = logger.histogram("histogram/histogram0", num_buckets=50)
+    histogram1 = logger.histogram("histogram/histogram1", num_buckets=50)
+
+```
+
+然后在Keras提供的回调函数（callback）中插入我们的数据采集代码就可以了。
+
+```python
+train_step = 0
+
+class LossHistory(keras.callbacks.Callback):
+    def on_train_begin(self, logs={}):
+        self.losses = []
+
+    def on_batch_end(self, batch, logs={}):
+        global train_step
+
+        # Scalar
+        scalar_keras_train_loss.add_record(train_step, logs.get('loss'))
+
+        # get weights for 2 layers
+        W0 = model.layers[0].get_weights()[0] # 3 x 3 x 1 x 32
+        W1 = model.layers[1].get_weights()[0] # 3 x 3 x 32 x 64
+
+        weight_array0 = W0.flatten()
+        weight_array1 = W1.flatten()
+
+        # histogram
+        histogram0.add_record(train_step, weight_array0)
+        histogram1.add_record(train_step, weight_array1)
+
+        # image
+        image0.start_sampling()
+        image0.add_sample([9, 32], weight_array0)
+        image0.finish_sampling()
+
+        image1.start_sampling()
+        image1.add_sample([288, 64], weight_array1)
+        image1.finish_sampling()
+
+        train_step += 1
+        self.losses.append(logs.get('loss'))
+```
+
+训练结束后，各个组件的可视化结果如下：
+
+关于误差的数值图的如下：
+
+<p align=center>
+<img width="70%" src="https://github.com/daming-lu/large_files/blob/master/keras_demo_figs/keras_scalar.png?raw=true" />
+</p>
+
+训练过后的第一，第二层卷积权重图的如下：
+
+<p align=center>
+<img width="70%" src="https://github.com/daming-lu/large_files/blob/master/keras_demo_figs/keras_image.png?raw=true" />
+</p>
+
+
+训练参数的柱状图的如下：
+
+<p align=center>
+<img width="70%" src="https://github.com/daming-lu/large_files/blob/master/keras_demo_figs/keras_histogram.png?raw=true" />
+</p>
+
+
+
+完整的演示程序可以在[这里](./keras_mnist_demo.py)下载。

demo/keras/keras_mnist_demo.py

@@ -0,0 +1,120 @@
+from __future__ import print_function
+import keras
+from keras.datasets import mnist
+from keras.models import Sequential
+from keras.layers import Dense, Dropout, Flatten
+from keras.layers import Conv2D, MaxPooling2D
+from keras import backend as K
+
+from visualdl import LogWriter
+
+batch_size = 2000
+num_classes = 10
+epochs = 10
+
+# input image dimensions
+img_rows, img_cols = 28, 28
+
+# the data, shuffled and split between train and test sets
+(x_train, y_train), (x_test, y_test) = mnist.load_data()
+
+if K.image_data_format() == 'channels_first':
+    x_train = x_train.reshape(x_train.shape[0], 1, img_rows, img_cols)
+    x_test = x_test.reshape(x_test.shape[0], 1, img_rows, img_cols)
+    input_shape = (1, img_rows, img_cols)
+else:
+    x_train = x_train.reshape(x_train.shape[0], img_rows, img_cols, 1)
+    x_test = x_test.reshape(x_test.shape[0], img_rows, img_cols, 1)
+    input_shape = (img_rows, img_cols, 1)
+
+x_train = x_train.astype('float32')
+x_test = x_test.astype('float32')
+x_train /= 255
+x_test /= 255
+print('x_train shape:', x_train.shape)
+print(x_train.shape[0], 'train samples')
+print(x_test.shape[0], 'test samples')
+
+# convert class vectors to binary class matrices
+y_train = keras.utils.to_categorical(y_train, num_classes)
+y_test = keras.utils.to_categorical(y_test, num_classes)
+
+model = Sequential()
+model.add(
+    Conv2D(32, kernel_size=(3, 3), activation='relu', input_shape=input_shape))
+model.add(Conv2D(64, (3, 3), activation='relu'))
+model.add(MaxPooling2D(pool_size=(2, 2)))
+model.add(Dropout(0.25))
+model.add(Flatten())
+model.add(Dense(128, activation='relu'))
+model.add(Dropout(0.5))
+model.add(Dense(num_classes, activation='softmax'))
+
+model.compile(
+    loss=keras.losses.categorical_crossentropy,
+    optimizer=keras.optimizers.Adadelta(),
+    metrics=['accuracy'])
+
+# create VisualDL logger
+logdir = "/workspace"
+logger = LogWriter(logdir, sync_cycle=100)
+
+# mark the components with 'train' label.
+with logger.mode("train"):
+    # create a scalar component called 'scalars/'
+    scalar_keras_train_loss = logger.scalar("scalars/scalar_keras_train_loss")
+    image0 = logger.image("images/image0", 1)
+    image1 = logger.image("images/image1", 1)
+    histogram0 = logger.histogram("histogram/histogram0", num_buckets=50)
+    histogram1 = logger.histogram("histogram/histogram1", num_buckets=50)
+
+train_step = 0
+
+
+class LossHistory(keras.callbacks.Callback):
+    def on_train_begin(self, logs={}):
+        self.losses = []
+
+    def on_batch_end(self, batch, logs={}):
+        global train_step
+
+        # Scalar
+        scalar_keras_train_loss.add_record(train_step, logs.get('loss'))
+
+        # get weights for 2 layers
+        W0 = model.layers[0].get_weights()[0]  # 3 x 3 x 1 x 32
+        W1 = model.layers[1].get_weights()[0]  # 3 x 3 x 32 x 64
+
+        weight_array0 = W0.flatten()
+        weight_array1 = W1.flatten()
+
+        # histogram
+        histogram0.add_record(train_step, weight_array0)
+        histogram1.add_record(train_step, weight_array1)
+
+        # image
+        image0.start_sampling()
+        image0.add_sample([9, 32], weight_array0)
+        image0.finish_sampling()
+
+        image1.start_sampling()
+        image1.add_sample([288, 64], weight_array1)
+        image1.finish_sampling()
+
+        train_step += 1
+        self.losses.append(logs.get('loss'))
+
+
+history = LossHistory()
+
+model.fit(
+    x_train,
+    y_train,
+    batch_size=batch_size,
+    epochs=epochs,
+    verbose=1,
+    validation_data=(x_test, y_test),
+    callbacks=[history])
+score = model.evaluate(x_test, y_test, verbose=0)
+print('Test loss:', score[0])
+print('Test accuracy:', score[1])