In the load test, the trained model present such wrong
Tensorflow. Python. Framework. Errors_impl. NotFoundError: Restoring the from checkpoint failed. This is most likely due to a Variable name or other graph key that is missing from the checkpoint. Both Please ensure that you have not altered the graph expected -based on the checkpoint, the Original error:

Key pose_net/Stage2/unit_1 depthwise/depthwise_weights not found in:
[[node save/RestoreV2 (defined at test_kitti_pose. Py: 65)]]
I used a tf in network structure. The nn. Depthwise_conv2d
The function of the weight I have defined the variable
W=tf. Get_variable (
'depthwise_weights'
[the kernel, the kernel, in_channels, 1], dtype=tf. Float32,
, initializer=weights_initializer
)
Started after the print can be training with the parameters in variable pose_net/Stage2/unit_1/depthwise/depthwise_weights, but nothing back view node file, this is what reason, have a great god solve, thank you very much!!