ILG2021
ILG2021
These days, I am working on a ASR software for filipino, but the results always contains grammer errors. I need a gec dataset to train a t5 gec model for...
I use whisper-jax in TPU for my speech translate project, I found that when it recognize a long sentence, it will cause 40s. But faster_whisper with a kaggle t4 only...
1. add main.py for the example 2. add support for chinese 3. add support custom faster whisper model
I request this feature because sometimes I will cut a part the audio, then all the labels after the part will miss position.