StuartIanNaylor

@robin1001 Hows things going? I have been searching for some form of binaural audio source seperation alg/model and there are some but seem to be offline than online/realtime. If you...

I have seen your repo's @JusperLee and second they also world be a great option and hope you do. I still think multi-channel help more with reverberation and often post...

https://github.com/BUTSpeechFIT/speakerbeam There was this paper that looks really interesting but can not find any code examples https://arxiv.org/abs/2106.02934 Also with Esspresif's laterial thought on BSS you could replace the KWS on...

sudo nano configure.py Change 2 lines to the following:- def is_raspi_arm(): return (os.uname()[4] == "armv7l") or (os.uname()[4] == "aarch64") if is_macos() or is_linux(): # write("build --copt=-mavx")

Any chance we could get this pull request merged as it doesn't actually change existing just adds support for missing architecture?

Hi, Robin. My personal feeling is the KWS we have are good but the manner we train and the datasets we use are not so great. KWS are usually simple...

I just noticed Mining Effective Negative Training Samples for Keyword Spotting ([github](https://github.com/jingyonghou/KWS_Max-pooling_RHE), [paper](https://www.microsoft.com/en-us/research/uploads/prod/2020/04/ICASSP2020_Max_pooling_KWS.pdf)) I have been wondering about a dataset creator and how to select !KW without class imbalance. I...

torch.cuda.mem_get_info() (7319715840, 8331132928) With an 8gb Quadro so not so much

Running the tflite runtime the quantised models are much faster on the intended Pi hardware.

I guess its down to destination more than anything as with many micro-controllers and npus recurrent layers often don't have support. I am not sure if tf4micro does or not...