FunASR icon indicating copy to clipboard operation
FunASR copied to clipboard

A Fundamental End-to-End Speech Recognition Toolkit and Open Source SOTA Pretrained Models, Supporting Speech Recognition, Voice Activity Detection, Text Post-processing etc.

Results 555 FunASR issues
Sort by recently updated
recently updated
newest added
trafficstars

下载了一个闽南语ASR预训练模型:https://www.modelscope.cn/models/chenyongxian299/speech_UniASR_asr_2pass-minnan-16k-common-vocab3825 由于它是pytorch模型格式,不能用funasr_wss_server 加载。需要先转为onnx格式。 于是参照项目的readme: `funasr-export ++model=speech_UniASR_asr_2pass-minnan-16k-common-vocab3825 ++quantize=false ++device=cpu` 但是执行报错: AssertionError: speech_UniASR_asr_2pass-minnan-16k-common-vocab3825 is not registered 请问怎么解决呢?

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) ## ❓ Questions and Help ### Before asking: 1. search the issues. 2. search the...

question

1.我将麦克风的流式音频转换成chunk输入模型,但是识别的效果比wav的本地音频要差很多,同时有很多误识别,请问这个是什么原因,加上一些vad会好吗? 2.还有,这个paraformer-zh是运行在gpu上面的吗?我看显存确实有相应增加?但是read me中似乎写着gpu未实现? sd.default.device = 27 # ID为27号设备 1 model = AutoModel(model="paraformer-zh-streaming") chunk_size = [0, 20, 5] encoder_chunk_look_back = 4 # number of chunks to lookback for encoder self-attention decoder_chunk_look_back...

question

## 🐛 Bug **FunASR Docker 镜像在 x86_64 架构上无法下载 - 特定层损坏** ### To Reproduce Steps to reproduce the behavior (**always include the command you ran**): 1. Run cmd 'docker pull registry.cn-hangzhou.aliyuncs.com/funasr_repo/funasr:funasr-runtime-sdk-cpu-0.4.7'...

bug

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) ## ❓ Questions and Help ### Before asking: 1. search the issues. 2. search the...

question

当准备的jsonl 数据中 路径含有 "aishell", 在准备训练数据时会将target 中的空格全部过滤掉, 导致英文文本空格全部丢失, 具体代码在funasr/datasets/audio_datasets/indes_ds.py 第96行 if "aishell" in source: target = target.replace(" ", "") # 可以再加一个关于text_language 是 "zh" 的判断 当有英文训练数据时, 可以修改路径名称规避掉这个问题

bug

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) ## ❓ Questions and Help ``` import logging, os, tqdm # 1. 让 ModelScope 闭嘴...

question

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 2pass模式时出现内存泄漏,泄漏速度取决于音频大小 ### To Reproduce Steps to reproduce the behavior (**always include the...

bug

1、**变长模型推理**:使用模型输入为(1,feats_length,400)的模型。流式推理,step=5120(320ms,16k采样率)。 ``` step = 5120 # 步长,5120 = 0.32s*16000,时长(s)*采样率 param_dict = {"in_cache": []} vad_segments = [] # 存储VAD检测结果 for sample_offset in range(0, speech_length, step): end = min(sample_offset + step, speech_length)...

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) ## ❓ Questions and Help 有2个问题 1. 我已经下载了模型文件在本地,为什么每次运行模型时,还会提示下载,这个下载步骤可以关掉吗? 2. 执行过程中如图所示的输出可以禁用掉吗? ### Before asking: 1. search...

question