-
Notifications
You must be signed in to change notification settings - Fork 1.5k
Description
Notice: In order to resolve issues more efficiently, please raise issue following 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)
🐛 Bug
我使用官方提供的代码例子:FunASR/runtime/python/websocket/funasr_wss_server.py
将model_asr模型换成 FunAudioLLM/Fun-ASR-Nano-2512,
出现报错:
UnboundLocalError: local variable 'AutoTokenizer' referenced before assignment
报错截图如下:
To Reproduce
Code sample
funasr_wss_server.py 修改了一些内容如下:
model_asr = AutoModel(
model=args.asr_model,
model_revision=args.asr_model_revision,
ngpu=args.ngpu,
ncpu=args.ncpu,
device=args.device,
disable_pbar=True,
disable_log=True,
trust_remote_code=True, ## 添加内容
remote_code="../../../funasr/models/fun_asr_nano/model.py", ## 添加内容
)
model_vad = AutoModel(
model=args.vad_model,
model_revision=args.vad_model_revision,
ngpu=args.ngpu,
ncpu=args.ncpu,
device=args.device,
disable_pbar=True,
disable_log=True,
trust_remote_code=True, ## 添加内容
remote_code="../../../funasr/models/fun_asr_nano/model.py", ## 添加内容
# chunk_size=60,
)
其中 trust_remote_code 和 remote_code无论是否添加,都会出现同样的报错
运行脚本如下:
export MODELSPATH=/data/ASR/FunASR/models/models/iic
python funasr_wss_server.py --port 10095
--asr_model FunAudioLLM/Fun-ASR-Nano-2512
--asr_model_revision master
--asr_model_online $MODELSPATH/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-online
--asr_model_online_revision v2.0.4
--vad_model $MODELSPATH/speech_fsmn_vad_zh-cn-16k-common-pytorch
--vad_model_revision v2.0.4
--punc_model $MODELSPATH/punc_ct-transformer_zh-cn-common-vad_realtime-vocab272727
--punc_model_revision v2.0.4
--ngpu 1
--ncpu 4
--certfile ''
--keyfile /data/ASR/FunASR/runtime/ssl_key/server.key \
Environment
- OS : Linux
- FunASR Version 1.2.9
- ModelScope Version : 1.32.0
- PyTorch Version : 2.9.1
- How you installed funasr: pip
- Python version: 3.10.19
- GPU : 3090
- CUDA/cuDNN version: 12.2
- Docker version (e.g., funasr-runtime-sdk-cpu-0.4.1) gpu版本,自制