FireRedTeam

Results 58 comments of FireRedTeam

fix torch error

Thanks for your pull request! We'll review the code.

请问默认的解码参数是什么

可以默认使用inference_fireredasr_aed.sh和inference_fireredasr_llm.sh中的解码参数

What is the best practice for extra long audio?

1. use VAD to split long wav, each part shorter than 60s (FireRedASR-AED) or 30s (FireRedASR-LLM) 2. use FireRedASR to do ASR

Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint.

https://github.com/FireRedTeam/FireRedASR/pull/14#issuecomment-2652872571

FireRedASR-AED-XS/S/M的版本大佬们方便开源嘛

内部流程中

[Feature Request] How to use the models to process long audio files?

Thanks for you request :) We will discuss this internally. Maybe you can use other VAD models to cut long audio files right now.

请问对粤语的支持怎么样？

没具体测过。但是模型应该是有一些粤语歌曲+粤语能力的。有兴趣可以找个测试集测一下，能反馈结果的话~提前感谢

Benchmark with gpt-4o-transcribe

What is your test set?

Is any manual about training on other language?

maybe a future work :)

方言识别问题

不需要