End-to-end-ASR-Pytorch
End-to-end-ASR-Pytorch copied to clipboard
This is an open source project (formerly named Listen, Attend and Spell - PyTorch Implementation) for end-to-end ASR implemented with Pytorch, the well known deep learning toolkit.
I didn't see ADVERSARIAL TRAINING. Can anyone tell me?
Our lab's GPU0 is used, and I can only use GPU1. It shows CUDA out of memory when using GPU0
请问 ctc.py 中的 cheap_compute 方法没有看明白,可不可以提供些相关资料
I want to test this model on a Windows laptop
Where to define or provide characters of non-English language? Or what modifications are required to apply this repo to non-English language
can i get words alignment from attention map to get word timestamp for each word in hypothisis
Can the audio duration be as long as two minutes?
when i test on aishell1 dataset, the WER is always 1.0, no drop