FireRedASR icon indicating copy to clipboard operation
FireRedASR copied to clipboard

FireRedASR-LLM报错:index out of range in self

Open extraordinaryW opened this issue 9 months ago • 4 comments

原始输入[[{'role': 'user', 'content': '请转写音频为文字'}, {'role': 'assistant', 'content': ''}]]转换为token_id后得到 [[151644, 872, 198, 151646, 14880, 46670, 61443, 111268, 17714, 87335, 151645, 198, 151644, 77091, 198]]其中“speech”的token_id为"151646" 而在embedding函数中打印weight形状为:torch.Size([151646, 3584]) 发生报错“index out of range in self”,为什么会发生这种情况?

Image

extraordinaryW avatar Mar 06 '25 04:03 extraordinaryW

解决了吗?

FireRedTeam avatar Mar 10 '25 04:03 FireRedTeam

解决了吗?

还没有,求方案。。

extraordinaryW avatar Mar 10 '25 09:03 extraordinaryW

你是直接跑inference_fireredasr_llm.sh报错了,还是你自己写了点python代码,报错了?

FireRedTeam avatar Mar 11 '25 04:03 FireRedTeam

是在直接使用项目自带的wav,运行 inference_fireredasr_llm.sh的时候就提示了上述错误

---Original--- From: @.> Sent at: 2025年3月11日(Tue) Noon12:05 To: @.>; Cc: @.@.>; Subject: Re: [FireRedTeam/FireRedASR] FireRedASR-LLM报错:index out of range in self (Issue #43)

你是直接跑inference_fireredasr_llm.sh报错了,还是你自己写了点python代码,报错了?

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***> FireRedTeam left a comment (FireRedTeam/FireRedASR#43)

你是直接跑inference_fireredasr_llm.sh报错了,还是你自己写了点python代码,报错了?

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

extraordinaryW avatar Mar 11 '25 04:03 extraordinaryW