如何将 TensorVariable 转换为 numpy

Question

我想将 TensorVariable 转换为 numpy 数组并尝试：

feature_vector = keras_model.get_layer(blob_name).output.numpy()

但是得到错误。

AttributeError: 'TensorVariable' object has no attribute 'numpy'

我也试过：

feature_vector = keras_model.get_layer(blob_name).output
init = tf.compat.v1.global_variables_initializer()

with tf.compat.v1.Session() as sess:
    sess.run(init)
    print(feature_vector.eval())

但是报错

theano.gof.fg.MissingInputError: Input 0 of the graph (indices start from 0), used to compute Shape(/input_1), was not provided and not given a value. Use the Theano flag exception_verbosity='high', for more information on this error.

Answer 1

谢谢@Lau。是的，我使用 theano 并修复了这个错误，如 this

如何将 TensorVariable 转换为 numpy

How to convert TensorVariable to numpy

numpy

keras

tensorflow

keras-layer

tensor