Implementation Status

[faroit]

The nnabla release of open-unmix is not feature complete with respect to our pytorch reference. The following issues would need to be solved

• [X] Dataset parameters
• [ ] Training parameters
• [ ] Inference
• [ ] Training/Inference on CPU (currently the nnabla LSTM does rely on cudnn)
• [ ] Evaluation

[mehdiziouziou504]

How can I listen music on open unmix ?