Alexey Korepanov comments

Results 43 comments of


                                            Alexey Korepanov

I've made a torch reimplementation for both offline and streaming implementation. Would you be interested in accepting this contribution?

@zeynepgulhanuslu I understood the problem. Jitsi if im not mistaken run wasm module inside AudioWorklet. So there you have only `128 / sampleRate * 1000` window (here is 2.6 ms)....

I've made a torch reimplementation for both offline and streaming implementation. Would you be interested in accepting this contribution?

I think you can remove normalization steps for model (in your cpp code). I don't do it and it works fine. Also you can profile your solution in browser. I...

I've made a torch reimplementation for both offline and streaming implementation. Would you be interested in accepting this contribution?

@GreatDarrenSun I'm not currently working on this because the current latency suits my current needs. But maybe someday later I'll start

I've made a torch reimplementation for both offline and streaming implementation. Would you be interested in accepting this contribution?

@308627993 You can find it here - https://github.com/Rikorose/DeepFilterNet/blob/5224ed47e19b6b327c9f22df6c5a061e3c3f6d5f/torchDF/torch_df_streaming.py Here is PR - https://github.com/Rikorose/DeepFilterNet/pull/433/

I've made a torch reimplementation for both offline and streaming implementation. Would you be interested in accepting this contribution?

Well, you can try infer with `apply_all_stages=True` parameter (in init), so you will get output like in offline model. I bet problem here. It's made to reach full match with...

I've made a torch reimplementation for both offline and streaming implementation. Would you be interested in accepting this contribution?

Can you send an audio please? I'll try to debug it tomorrow

Pure torch implementation

@dingchaoyue here is original code, you can calculate feature bank using it https://github.com/Rikorose/DeepFilterNet/blob/59789e135cb5ed0eb86bb50e8f1be09f60859d5c/libDF/src/lib.rs#L62

Pure torch implementation

1. I'm using the pyDF changes to run a test to compare the original Rust and Torch threading implementations. What do you think about the tests then? Do I need...

Pure torch implementation

@Rikorose I saw your comment, right not I have more important tasks now. But I'll come back in a while and fix it.

Pure torch implementation

Hello! Can you provide more details? What code are you running? Encoder in this reimplementation takes 4 parameters https://github.com/grazder/DeepFilterNet/blob/1097015d53ced78fb234e7d7071a5dd4446e3952/DeepFilterNet/df/deepfilternet3.py#L168 https://github.com/grazder/DeepFilterNet/blob/1097015d53ced78fb234e7d7071a5dd4446e3952/torchDF/torch_df_streaming.py#L492 ``` self, feat_erb: Tensor, feat_spec: Tensor, hidden: Tensor ``` Btw...