QuantumNous / QuantumNous/new-api

无法调用Groq渠道的Whisper模型处理 .wav 音频文件,报错“令牌无权访问模型”

Open
#952 5 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

bug stale
Dominant language
Go
Stars
48.5k
Forks
11.6k
Avg merge
1d 17h
Merged PRs (30d)
58

Description

例行检查

  • 我已确认目前没有类似 issue
  • 我已确认我已升级到最新版本
  • 我已完整查看过项目 README,尤其是常见问题部分
  • 我理解并愿意跟进此 issue,协助测试和提供反馈
  • 我理解并认可上述内容,并理解项目维护者精力有限,不遵循规则的 issue 可能会被无视或直接关闭

问题描述
在new-api中添加groq渠道,测试groq的whisper模型,发现对于 .wav的音频文件无法调用模型。
报错{"error":{"message":"该令牌无权访问模型 whisper-large-v3-turbo (request id: 20250410005049655360160lkh0pltW)","type":"new_api_error"}}

无论是使用"Content-Type: multipart/form-data" ,失败

curl http://127.0.0.1:3031/v1/audio/transcriptions \
  -H "Authorization: Bearer sk-4tv7Of7QJJVWA1fKmYHYkGyRWYOeD5c373fXoSGZkcAgECYp" \
  -H "Content-Type: multipart/form-data" \
  -F file="@/Users/bulejames/Downloads/20250410 0034 record-yourself.wav" \
  -F model=whisper-large-v3-turbo
{"error":{"message":"该令牌无权访问模型 whisper-large-v3-turbo (request id: 20250410005049655360160lkh0pltW)","type":"new_api_error"}}

还是 "Content-Type: audio/wav",失败

curl http://127.0.0.1:3031/v1/audio/transcriptions \
  -H "Authorization: Bearer sk-4tv7Of7QJJVWA1fKmYHYkGyRWYOeD5c373fXoSGZkcAgECYp" \
  -H "Content-Type: audio/wav" \
  -F file="@/Users/bulejames/Downloads/20250410 0034 record-yourself.wav" \
  -F model=whisper-large-v3-turbo       
{"error":{"message":"该令牌无权访问模型 whisper-1 (request id: 20250410004936408826757hPnZTt30)","type":"new_api_error"}}

但是使用 .mp3格式的音频文件就可以正常调用

curl http://127.0.0.1:3031/v1/audio/transcriptions \
  -H "Authorization: Bearer sk-a1LnjRU5RSVWEk9Z8scvmO0iYlOHWHSgCwyUu0FgHcryFNq7" \
  -H "Content-Type: multipart/form-data" \
  -F file="@/Users/bulejames/Downloads/20250104 1128 Recording.mp3" \     
  -F model=whisper               
{"text":" I saw her at the meeting.","x_groq":{"id":"req_01jrdqb5jcfxqsmrtrqkdn78s5"}}

复现步骤

shell命令

curl http://127.0.0.1:3031/v1/audio/transcriptions \
  -H "Authorization: Bearer sk-4tv7Of7QJJVWA1fKmYHYkGyRWYOeD5c373fXoSGZkcAgECYp" \
  -H "Content-Type: audio/wav" \
  -F file="@/Users/bulejames/Downloads/20250410 0034 record-yourself.wav" \
  -F model=whisper-large-v3-turbo       

预期结果

正常调用groq的whisper模型转录成功

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Run the reported curl reproduction against /v1/audio/transcriptions and inspect the request path handling Groq Whisper uploads. Compare the WAV and MP3 behavior and model selection; done means a WAV upload reaches Groq and returns a transcription without the token or model-access error.

Written by the indexing model from the issue text.

Assessment

Tech stack
go
Domain
api, audio-video-rtc
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Quiet
Clarity
Needs clarification
Newbie friendliness
42/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.