File "/home/wjs/.local/lib/python3.7/site-packages/python/cifar10/cifar10_train.py", line 69, in train
images, labels = cifar10.distorted_inputs()
File "/home/wjs/.local/lib/python3.7/site-packages/python/cifar10/cifar10.py", line 151, in distorted_inputs
batch_size=FLAGS.batch_size)
File "/home/wjs/.local/lib/python3.7/site-packages/python/cifar10/cifar10_input.py", line 164, in distorted_inputs
distorted_image = tf.image.random_crop(reshaped_image, [height, width])
File "/home/wjs/.local/lib/python3.7/site-packages/tensorflow/python/ops/random_ops.py", line 317, in random_crop
math_ops.reduce_all(shape >= size),
File "/home/wjs/.local/lib/python3.7/site-packages/tensorflow/python/ops/gen_math_ops.py", line 4271, in greater_equal
"GreaterEqual", x=x, y=y, name=name)
ValueError: Dimensions must be equal, but are 3 and 2 for 'random_crop/GreaterEqual' (op: 'GreaterEqual') with input shapes: [3], [2].
求问应该修改哪里的参数。
问题已解决。
解决方案是在cifar-10 input里改变random_crop参数,如下所示:
distorted_image = tf.image.random_crop(reshaped_image, [height, width,3])