mirror of
https://github.com/k2-fsa/icefall.git
synced 2025-08-08 17:42:21 +00:00
3e4da5f781
* Replace deprecated pytorch methods
- torch.cuda.amp.GradScaler(...) => torch.amp.GradScaler("cuda", ...)
- torch.cuda.amp.autocast(...) => torch.amp.autocast("cuda", ...)
* Replace `with autocast(...)` with `with autocast("cuda", ...)`
Co-authored-by: Li Peng <lipeng@unisound.ai>
Introduction
This recipe includes scripts for training Qwen-Audio style model using multiple datasets.
./RESULTS.md contains the latest results.
ASR_LLM
The following table lists the folders for different tasks.
| Speech Encoder | LLM | Comment | |
|---|---|---|---|
| whisper_llm_zh | Whisper | Qwen2 | Using multiple Chinese datasets |