SiyuanWei comments

Results 8 comments of


                                            SiyuanWei

Mutli gpu training

I have a image computing PC with cpu(16g) and gpu(8g) ，is it enough to train the model without throwing a CUDA out of Memory error,please?

Mutli gpu training

thanks anyway ,although it is a bad news

SETraining is extremely slow

And I cProfile the SE script and it shows the pykaldi asr decode funtion is the bottleneck

SETraining is extremely slow

I use 8 batch size and decoding graph and transition model under a tri4b system of kaldi librispeech examples. Other configurations shouldn't infulence the decoding speed...emm

Question about the TPS

Sorry for the late reply, I didn’t check my GitHub in time. For dTPS, we use gumbel-softmax to help train learnable token scoring head ,which follows dynamicViT. So, in the...

Sorry for the late reply, I didn’t check my GitHub in time. basecls can be installed by following the instruction from [https://github.com/megvii-research/basecls], "pip install " should be enough. Have you...

Hope the whole source code can be released

Thanks ! We are working on this !

Hope the whole source code can be released

Sorry for the late reply. The codes of our eTPS implemented by megengine can be found in https://github.com/megvii-research/TPS-CVPR2023/blob/main/megengine_codebase/models/tps.py#L91. Sorry that we haven't released the whole implementation of eTPS on the...