Anybody get train_multi_gpus.py and eval.py working?

[cozec]

Trained with 8GPU and see error msg below. This never happens to single GPU workflow.

NotFoundError (see above for traceback): Key net/encoder/norm1/Variable not found in checkpoint [[Node: save/RestoreV2_115 = RestoreV2[dtypes=[DT_FLOAT], _device="/job:localhost/replica:0/task:0/cpu:0"](_arg_save/Const_0_0, save/RestoreV2_115/tensor_names, save/RestoreV2_115/shape_and_slices)]]

[qclu]

I tried with train_multi_gpus.py, and traning is fine. But with eval.py, I get same errores. @cozec Have u solved this?

[cozec]

not yet. busy with training using phoneme