modelscope/FunASR
A Fundamental End-to-End Speech Recognition Toolkit and Open Source SOTA Pretrained Models, Supporting Speech Recognition, Voice Activity Detection, Text Post-processing etc.
PythonNOASSERTION
Issues
- 5
- 2
docker拉取镜像失败,只能拉取部分,导致服务端运行失败
#2037 opened by Xuyaoyan - 5
能不能自己打镜像?能否提供Dockerfile
#2010 opened by MrRohwei - 1
runtime/csharp/ws-client/FunASRWSClient_Offline 项目不支持并发
#2013 opened by dfengpo - 0
wrong submit. sorry for that.
#2023 opened by laishuzhong - 2
GPU版本:[W shape_analysis.cpp:841] failed PropagateTensorShapeOnNode with schema
#2021 opened by zhanglv0209 - 1
SenseVoiceSmall没有带时间戳
#2027 opened by SpringSan - 1
docker部署funasr,镜像是funasr-runtime-sdk-cpu-0.4.5,如何识别不同人说话
#2033 opened by jaychuo - 1
当我通过run_server.sh使用SenseVoiceSmall模型时出现错误:models/iic/SenseVoiceSmall/model_quant.onnx do not exists.
#2035 opened by firekirin67 - 1
A10卡GPU推理效率和CPU持平,不清楚是什么地方的问题
#2042 opened by lanyuer - 1
docker部署,wave.Error: unknown format: 6,ffmpeg没有如期运行
#2044 opened by Jack-Lin-gif - 1
离线文件服务websocket解析音频后,返回的结果不对
#2054 opened by wangguo1230 - 2
微调seaco_paraformer模型单机多卡出现错误
#2041 opened by gzqqqqqq - 1
现在这个FunASR能用在普通的商业上吗?
#2031 opened by dyjiangjh - 3
docker gpu1.1长文件时间戳偏移很大
#2019 opened by dfengpo - 2
基于paraformer & whisper 的离线转写网页,时间戳断句试听,生成字幕等
#2059 opened by pika-online - 0
实时转写docker服务是否支持发送message参数设置是否添加标点符号
#2045 opened by yeyupiaoling - 2
"emotion2vec_plus_large" model does not exist
#2050 opened by Rob-Nafous - 1
多段语音并行识别失败问题
#2051 opened by MENGHAH - 1
- 1
ValueError: math domain error
#2060 opened by lsdlh - 1
音频开始有个类似开幕音,会被识别为“嗯”。这需要如何配置,还是需要微调
#2061 opened by flystar8 - 1
尝试使用 finetune.sh 微调 seaco-paraformer ,微调后发现,每个epoch 的model.pt增大了很大,从原模型的800多M,增大到了2.44GB多,这是什么原因
#2032 opened by gzqqqqqq - 1
- 3
实时转写 Docker 服务如何使用讲话人识别?
#2048 opened by purerosefallen - 3
- 3
funasr-runtime-sdk-gpu-0.1.1版本运行过程中出现异常
#2058 opened by liurongjie174 - 1
pip insatll funasr 以后,logging 不能正常打印
#2036 opened by dibaotian - 4
- 0
label_smoothing_loss target None
#2020 opened by kli017 - 0
不行,windows11 on arm64 电脑上运行不了
#2056 opened by wmx-github - 0
speech_ngram_lm_zh-cn-ai-wesp-fst 这个模型怎么做增量训练
#2052 opened by liuwenchang - 0
文档写的真多,过多了
#2049 opened by Noah0115 - 0
FunASR版本的Paraformer模型什么什么时候能支持音频中的数字转译?
#2047 opened by smengfei - 0
c++ runtime 2pass 循环解码同一句话,会使得offline模型解码越来越慢吗?
#2046 opened by locasxe - 1
加入spk_model后,一个模型似乎只能推理一次
#2039 opened by orangenuinee - 0
How to chose quantizing value when exporting? 16 bit vs 8 bit (for streaming)
#2043 opened by andreystarenky - 1
ONNX Export of Pretrained models outputs garbage
#2040 opened by andreystarenky - 0
filename too long
#2038 opened by hjj-lmx - 4
merge_vad 后,推理遇到的两个问题
#2029 opened by Joy-word - 1
- 0
麦克风说话的音频的保存问题
#2030 opened by lxb0425 - 0
这几个模型名到底都能填写什么?
#2026 opened by leavegee - 0
parformer大规模数据集预训练cpu内存持续增加的问题
#2018 opened by chenpaopao - 2
无法理解传入的参数,使用ASR
#2017 opened by deadash - 0
请问vad模型是否支持 实时音频流的语音端点检测? 为什么每次检测音频活动端点都是'value': [[0, 1180]]}]
#2015 opened by viviliuwqhduhnwqihwqwudceygysjiwuwnn - 0
微调sensevoice模型Update best acc: 0.0000
#2016 opened by xiulianzw - 1
How to fine tune training without using GPU when using CPU version Docker images
#2014 opened by qqqq640 - 0
- 0
为什么finetune speech_seaco_paraformer_large_asr_nat-zh-cn-16k-common-vocab8404-pytorch,验证集Update best acc: 0.0000
#2012 opened by juzstu