tensorflow - Test a MNIST number -
after trained , evaluated model, trying feed model mnist test image :
the function logits fails. x_new_scaled 784 long array, scaled 0.01 0.99xxx method logits.eval expect reiceve? correct way pass array?
your question duplicate of typeerror: unhashable type: 'numpy.ndarray' tensorflow
the point use x twice:
x = mnist["data"] and
z = logits.eval(feed_dict={x: x_new_scaled}) so problem not x_new_scaled, problem key of feed_dict, being x not tensor a numpy.ndarray.

Comments
Post a Comment