python - Shape of a new variable must be fully defined, but instead was (?, 128) -
i training multi layer convolution network using tensorflow.i had no issues 2 layers.for third layer,when define weight,it gives me error shape of new variable must defined, instead (?, 128). saw sof link , used reuse=true, error persists.any helpful.
below code:
with tf.variable_scope('local3') scope: reshape = tf.reshape(pool2, shape=[batch_size, -1]) dim = reshape.get_shape()[1].value weights = tf.get_variable('weights', shape=[dim,128], dtype=tf.float32, initializer=tf.truncated_normal_initializer(stddev=0.005,dtype=tf.float32)) biases = tf.get_variable('biases', shape=[128], dtype=tf.float32, initializer=tf.constant_initializer(0.1)) local3 = tf.nn.relu(tf.matmul(reshape, weights) + biases, name=scope.name) here batch_size 32 , pool2 tensor of shape=(?, 14, 14, 16).
note:this code works when execute inside function.why so?
Comments
Post a Comment