Chengyi Wang
Chengyi Wang
I have met the same problem. Have you solved it?
Hi, I find the waveglow model generates nan tensors, which leads to the silent output. I fixed this issue by using fp32. You can try to remove .half() in load_waveglow...
> I have finished training and decoding in AISHELL-1 dataset and got cer=12.4% in test set，and i found that my model.json which uses the default config is different from the...
Sorry for the late reply. For streaming Conformer, you can simply set the conv module in the encoder to casual conv and the self-attention layers are as same as in...
The delay is computed by (chunk_size * downsample_rate * 10) + CNN_delay. So it should be (32 * 40) + 30
The future context depends on your right window size. For example, if each frame can only access the context within the same chunk (right_window=0), the future context is always 32...
Yes, you are right. I've fixed the bug in my repo now.