如何比较两组 TensorFlow 文件variables.index和variables.data-00000-of-00001？是否有任何实用程序可以让我以纯文本形式读取它们，或者以其他方式比较这些文件中的值？
为了查看存储在文件中的张量，只需执行以下操作：
from tensorflow.python.tools.inspect_checkpoint import print_tensors_in_checkpoint_file
print_tensors_in_checkpoint_file(file_name="model_name",tensor_name="",all_tensors=True)
输出将如下所示：
tensor_name: Decoder/decoder/Attention_Wrapper/attention_layer/kernel
[[-0.24866164 -0.0762616 0.02982535 ... 0.14274237 -0.0651149
0.18989304]
[ 0.16675466 0.01662545 0.00896074 ... -0.06836775 -0.02540337
0.0579135 ]
[-0.34853792 0.05886076 0.21691915 ... 0.03068756 0.37517232
0.0330036 ]
...
[-0.00463026 -0.05060914 -0.03661629 ... 0.03341271 -0.003078
0.04194186]
[ 0.07343228 -0.00445814 -0.01558546 ... -0.0164716 -0.00453405
0.0168137 ]
[ 0.03857246 -0.04361508 -0.04800353 ... 0.06200589 -0.05994397
-0.03183305]]
tensor_name: Decoder/decoder/Attention_Wrapper/attention_layer/kernel/training_op
[[1.7215452e-08 1.6511342e-08 2.3156689e-08 ... 6.9746142e-09
3.6705597e-08 6.0520193e-09]
[2.8541326e-06 9.9242357e-07 4.3421019e-07 ... 8.2636637e-07
2.9466917e-06 9.2266538e-07]
[1.4949121e-06 4.0124786e-07 1.8415257e-07 ... 2.1582312e-07
6.7197351e-07 3.1459973e-07]
...
[3.6029601e-15 1.2615142e-15 4.7981583e-15 ... 5.8665957e-16
2.0862790e-15 2.4295565e-15]
[5.8586951e-14 4.5164783e-14 2.6819809e-14 ... 4.2972126e-14
1.1934837e-13 3.0617263e-14]
[2.6801797e-15 9.8903675e-16 3.7704488e-16 ... 9.2021988e-16
2.9194079e-15 9.9945483e-16]]