modelscope / modelscope/FunASR
关于serve_realtime_ws.py服务的客户端测试达不到实时
Open
Nobody has claimed this yet.
benchmark
bug
deployment
needs feedback
question
- Dominant language
- Python
- Stars
- 20.4k
- Forks
- 2k
- Avg merge
- 4h 55m
- Merged PRs (30d)
- 169
Description
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。）
Question
我同时在浏览器中打开多份客户端测试demo，选择mic收音方式，在blibli上找一段长音频，播放进行测试，测试发现无法做到实时转写，播放的越久ASR识别延迟越大。
## Environment
- OS:
- Python version: 3.12.12
- FunASR version: 1.3.14
- ModelScope version:1.34.0
- PyTorch / torchaudio version: 2.10.0+cu129
- Install method (`pip`, source, Docker): pip
- Device (`cuda`, `cpu`, `mps`): L20 单卡
- vllm version: 0.16.0rc2.dev376+gf4af642a6
- CUDA/cuDNN version: 12.9
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reproducing the reported behavior with serve_realtime_ws.py and the browser client test demos, using microphone input and a long audio recording. Measure whether transcription latency grows during playback and compare the service behavior over time. Done means the client test remains real-time without accumulating delay.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- api, backend, performance, testing-qa
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Active
- Clarity
- Needs clarification
- Newbie friendliness
- 45/100