batch_size = 50
def my_eval(st,en):
x_test = mnist.test.images[st:en]
y_test = mnist.test.labels[st:en]
return accuracy.eval(feed_dict={x: x_test, y_: y_test, keep_prob: 1.0})
test_accuracy = np.mean([my_eval(i,i+batch_size) for i in range(0, mnist.test._num_examples, batch_size)])
--Verwenden Sie den Anfangswert von He
def weight_variable(shape):
"""weight_variable generates a weight variable of a given shape."""
if len(shape) == 4:
sd = np.sqrt(2.0/shape[0]/shape[1]/shape[2])
if len(shape) == 2:
sd = np.sqrt(2.0/shape[0])
initial = tf.truncated_normal(shape, stddev=sd)
return tf.Variable(initial)
train_step = tf.train.AdamOptimizer(1e-3).minimize(cross_entropy)
Recommended Posts