我正在尝试在Tensorflow中实现自己的自定义激活功能。
def leaky_relu_6(x):
if x >=0.0 and x < 6.0:
return x*1.0
elif x > 6.0:
return 6.0
else:
return 0.2 * x
np_leaky_relu_6= np.vectorize(leaky_relu_6)
def d_leaky_relu_6(x):
if x >=0.0 and x < 6.0:
return 1.0
elif x > 6.0:
return 0.0
else:
return 0.2
np_d_leaky_relu_6 = np.vectorize(d_leaky_relu_6)
def py_func(func, inp, Tout, stateful=True, name=None, grad=None):
# Need to generate a unique name to avoid duplicates:
rnd_name = 'PyFuncGrad' + str(np.random.randint(0, 1E+2))
tf.RegisterGradient(rnd_name)(grad) # see _MySquareGrad for grad example
g = tf.get_default_graph()
with g.gradient_override_map({"PyFunc": rnd_name}):
return tf.py_func(func, inp, Tout, stateful=stateful, name=name)
def relu_grad(op, grad):
x = op.inputs[0]
n_gr = tf_d_leaky_relu_6(x)
return grad * n_gr
np_leaky_relu_6_32 = lambda x: np_leaky_relu_6(x).astype(np.float32)
def tf_leaky_relu_6(x,name=None):
with tf.name_scope(name, "leaky_relu_6", [x]) as name:
y = py_func(np_leaky_relu_6_32,
[x],
[tf.float32],
name=name,
grad= relu_grad)
return y[0]
np_d_leaky_relu_6_32 = lambda x: np_d_leaky_relu_6(x).astype(np.float32)
def tf_d_leaky_relu_6(x,name=None):
with tf.name_scope(name, "d_leaky_relu_6", [x]) as name:
y = py_func(np_d_leaky_relu_6_32,
[x],
[tf.float32],
name=name,
stateful=False)
return y[0]
使用以下输入，代码可以正常运行。
with tf.Session() as sess:
x = tf.constant([0.2,0.7,1.2,-8.7])
y = tf_leaky_relu_6(x)
tf.initialize_all_variables().run()
print(x.eval(), y.eval(), tf.gradients(y, [x])[0].eval())
但是当我尝试将其与CNN代码一起使用进行分类时。
def cnn_model_fn(features, labels, mode):
"""Model function for CNN."""
# Input Layer
input_layer = tf.reshape(features["x"], [-1, 28, 28, 1])
conv1_act = tf.placeholder(tf.float32, [4,])
# Convolutional Layer #1
conv1 = tf.layers.conv2d(
inputs=input_layer,
filters=32,
kernel_size=[5, 5],
padding="same")
#activation=tf.nn.relu6)
#activation =tf_leaky_relu_6)
conv1_act = tf_leaky_relu_6(conv1)
print(tf.shape(conv1_act))
#output of the above print statement: Tensor("Shape:0", shape=(?,), dtype=int32)
pool1 = tf.layers.max_pooling2d(inputs=conv1_act, pool_size=[2, 2], strides=2)
# Convolutional Layer #2 and Pooling Layer #2
conv2 = tf.layers.conv2d(
inputs=pool1,
filters=64,
kernel_size=[5, 5],
padding="same",
activation=tf.nn.relu6) #to try other activation function replcae relu6 by leaky_relu or relu
pool2 = tf.layers.max_pooling2d(inputs=conv2, pool_size=[2, 2], strides=2)
# Dense Layer
pool2_flat = tf.reshape(pool2, [-1, 7 * 7 * 64])
dense = tf.layers.dense(inputs=pool2_flat, units=1024, activation=tf.nn.relu6) #to try other activation function replcae relu6 by leaky_relu or relu
dropout = tf.layers.dropout(
inputs=dense, rate=0.4, training=mode == tf.estimator.ModeKeys.TRAIN)
# Logits Layer
logits = tf.layers.dense(inputs=dropout, units=10)
predictions = {
# Generate predictions (for PREDICT and EVAL mode)
"classes": tf.argmax(input=logits, axis=1),
# Add `softmax_tensor` to the graph. It is used for PREDICT and by the
# `logging_hook`.
"probabilities": tf.nn.softmax(logits, name="softmax_tensor")
}
if mode == tf.estimator.ModeKeys.PREDICT:
return tf.estimator.EstimatorSpec(mode=mode, predictions=predictions)
# Calculate Loss (for both TRAIN and EVAL modes)
loss = tf.losses.sparse_softmax_cross_entropy(labels=labels, logits=logits)
# Configure the Training Op (for TRAIN mode)
if mode == tf.estimator.ModeKeys.TRAIN:
optimizer = tf.train.GradientDescentOptimizer(learning_rate=0.001)
train_op = optimizer.minimize(
loss=loss,
global_step=tf.train.get_global_step())
return tf.estimator.EstimatorSpec(mode=mode, loss=loss, train_op=train_op)
# Add evaluation metrics (for EVAL mode)
eval_metric_ops = {
"accuracy": tf.metrics.accuracy(
labels=labels, predictions=predictions["classes"])
}
return tf.estimator.EstimatorSpec(
mode=mode, loss=loss, eval_metric_ops=eval_metric_ops)
((train_data, train_labels),
(eval_data, eval_labels)) = tf.keras.datasets.mnist.load_data()
train_data = train_data/np.float32(255)
train_labels = train_labels.astype(np.int32) # not required
eval_data = eval_data/np.float32(255)
eval_labels = eval_labels.astype(np.int32) # not required
mnist_classifier = tf.estimator.Estimator(
model_fn=cnn_model_fn)
train_input_fn = tf.estimator.inputs.numpy_input_fn(
x={"x": train_data},
y=train_labels,
batch_size=100,
num_epochs=None,
shuffle=True)
# train one step and display the probabilties
#mnist_classifier.train(input_fn=train_input_fn, steps=1,hooks=[logging_hook])
mnist_classifier.train(input_fn=train_input_fn, steps=1,)
当我运行上述代码时，它给我以下错误：ValueError：max_pooling2d_1层的输入0与该层不兼容：其等级未定义，但该层需要定义的等级。
错误来自以下行：pool1