Skip to content

这模型是不是不适合用来生成字幕啊,精度不高,准确度不高,还有这哪几个模型识别字幕合适啊 #3799

Open

Description

@hothotds

Notice: In order to resolve issues efficiently, please follow the template.
(注意:为了更加高效率解决您遇到的问题,请按照模板提问,补充细节。)

Before asking

  1. Search existing issues: https://github.com/modelscope/FunASR/issues
  2. Search the docs: https://modelscope.github.io/FunASR/
  3. Check the README quick start and deployment section.

Question

Code or command

What have you tried?

Environment

  • OS:
  • Python version:
  • FunASR version:
  • ModelScope version:
  • PyTorch / torchaudio version:
  • Install method (pip, source, Docker):
  • Device (cuda, cpu, mps):
  • GPU model:
  • CUDA/cuDNN version:
  • Docker image tag, if used:

Activity

  1. added
    questionFurther information is requested
    needs triageNeeds maintainer triage and routing
    on Oct 27, 2026
  2. LauraGPT commented on Oct 28, 2026

    @LauraGPT
    Collaborator

    目前正文还是空白模板,仅凭标题还无法判断是文字识别错误,还是字幕分句、时间戳对齐的问题,也不能据此比较模型是否适合你的音频。

    请补充以下最小复现信息,方便针对同一段音频排查:

    • 实际使用的模型完整名称或路径、FunASR 版本和完整调用命令。
    • 音频语言、采样率、声道数,以及是否有背景音乐、多人重叠或很长的连续录音。
    • 一小段有权公开的原始音频(例如 10–30 秒),对应的期望文字与实际转写;若问题在字幕,请同时贴出对应 SRT 和预期时间范围。请勿上传未获授权或含敏感信息的音频。
    • 是 Python AutoModel、funasr-cli,还是服务端/其他工具生成字幕;是否经过 VAD、标点或时间戳处理。

    我们先把识别、分句和时间戳三个环节区分开,再讨论具体配置或模型选择。本 issue 保持打开,等待可复现信息。

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    needs triageNeeds maintainer triage and routingquestionFurther information is requested

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions