search

labring / FastGPT

FastGPT is a knowledge-based platform built on the LLMs, offers a comprehensive suite of out-of-the-box capabilities such as data processing, RAG retrieval, and visual AI workflow orchestration, letting you easily develop and deploy complex question-answering systems without the need for extensive setup or configuration.
https://tryfastgpt.ai
Other
17.41k stars 4.68k forks source link

为什么换了模型，还是走whisper #2047

Closed SDAIer closed 2 months ago

SDAIer commented 3 months ago

本地部署了whisper，接入onapi，fastgpt语音测试正常，只是识别效果很差。 发现了奇怪问题： 使用curl无论是向whisper 3003端口进行测试，还是向oneapi 3001端口测试时，model随便写什么都能输出文本内容，而且都是调用的whisper-1

而且我希望使用阿里的语音模型，日志显示一直走whisper

wihsper 3003端口 curl http://127.0.0.1:3003/v1/audio/transcriptions \ -H "Authorization: Bearer sk-tarzan" \ -H "Content-Type: multipart/form-data" \ -F model="paraformer-realtime-v1-XXXXXX" \ -F file="@/home/ftpzjp/1.mp3" \ -F response_format=text image

oneapi 3001端口 curl http://127.0.0.1:3001/v1/audio/transcriptions \ -H "Authorization: Bearer sk-fastgpt" \ -H "Content-Type: multipart/form-data" \ -F model="whisper-1" \ -F file="@/home/ftpzjp/1.mp3" \ -F response_format=text

image

config.json改了阿里是语音模型，oneapi在原来阿里渠道里增加了阿里语音模型，但是不生效，一直走whisper-1。 请问我哪里做的有问题吗

image

SDAIer commented 3 months ago

@songquanpeng

c121914yu commented 2 months ago

songquanpeng/one-api#1690

eric0095 commented 1 month ago

怎么解决这个问题呢？ 默认也是走whisper $ curl --request POST \ --url http://127.0.0.1:3001/v1/audio/transcriptions \ --header 'accept: application/json' \ --header 'authorization: Bearer aaa' \ --header 'content-type: multipart/form-data' \ --form file='@bbj.mp3' \ --form model=iic/SenseVoiceSmall % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 100 993k 100 132 100 993k 313 2359k --:--:-- --:--:-- --:--:-- 2359k{"error":{"message":"该令牌无权使用模型：whisper-1 (request id: 2024090310390341030820339103078)","type":"one_api_error"}}