From 6f334d49555bb3505fd7f782871b7f1f06c074d3 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E8=B5=B5=E6=B1=9F=E6=B1=9F?= Date: Thu, 22 May 2025 19:24:18 +0800 Subject: [PATCH 1/2] =?UTF-8?q?fix:=E4=BF=AE=E6=94=B9cosyvoice2=20README?= =?UTF-8?q?=E6=9D=83=E9=87=8D=E4=B8=8B=E8=BD=BD=E9=83=A8=E5=88=86=E5=86=85?= =?UTF-8?q?=E5=AE=B9?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .../audio/CosyVoice/CosyVoice2/README.md | 24 ++++++++++++------- 1 file changed, 16 insertions(+), 8 deletions(-) diff --git a/ACL_PyTorch/built-in/audio/CosyVoice/CosyVoice2/README.md b/ACL_PyTorch/built-in/audio/CosyVoice/CosyVoice2/README.md index 561641271a..31740a8f02 100755 --- a/ACL_PyTorch/built-in/audio/CosyVoice/CosyVoice2/README.md +++ b/ACL_PyTorch/built-in/audio/CosyVoice/CosyVoice2/README.md @@ -79,15 +79,23 @@ 4. 获取权重数据 - 本案例以CosyVoice2-0.5B为例,其他权重请自行适配 - - 获取 https://www.modelscope.cn/iic/CosyVoice2-0.5B 权重文件夹,放在CosyVoice目录下 + 本案例以CosyVoice2-0.5B为例,其他权重请自行适配。将下载下来的权重**放在CosyVoice目录下**。 + + 因cosyvoice2在2025年4月底更新过一次代码权重,因此需要使用`snapshot_download`下载指定commit id的权重。 + ```python + from modelscope.hub.snapshot_download import snapshot_download + + model_id = "模型名称" # 例如 "iic/CosyVoice2-0.5B" + commit_id = "9bd5b08fc085bd93d3f8edb16b67295606290350" + + model_dir = snapshot_download( + model_id, + revision=commit_id, # 关键,传入commit_id + cache_dir="./my_model" + ) + print(f"模型下载至:{model_dir}") + ``` - 或者通过git方式获取 - ``` - # git模型下载,请确保已安装git lfs - git clone https://www.modelscope.cn/iic/CosyVoice2-0.5B.git CosyVoice/CosyVoice2-0.5B - ``` 本用例采用sft预训练音色推理,请额外下载spk权重放到权重目录下 ``` wget https://www.modelscope.cn/models/iic/CosyVoice-300M-SFT/resolve/master/spk2info.pt -- Gitee From cebd6ea9aff7998583b7529b421ecc4f801605b7 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E8=B5=B5=E6=B1=9F=E6=B1=9F?= Date: Thu, 22 May 2025 20:01:42 +0800 Subject: [PATCH 2/2] =?UTF-8?q?fix:=E4=BF=AE=E6=94=B9sensevoice=20requirem?= =?UTF-8?q?ent=E6=96=87=E4=BB=B6?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- ACL_PyTorch/built-in/audio/SenseVoice/requirements.txt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/ACL_PyTorch/built-in/audio/SenseVoice/requirements.txt b/ACL_PyTorch/built-in/audio/SenseVoice/requirements.txt index 4a568e8a32..2a24970fa9 100755 --- a/ACL_PyTorch/built-in/audio/SenseVoice/requirements.txt +++ b/ACL_PyTorch/built-in/audio/SenseVoice/requirements.txt @@ -8,4 +8,4 @@ funasr>=1.1.13 numpy<=1.26.4 gradio fastapi>=0.111.1 -funars_onnx \ No newline at end of file +funasr_onnx \ No newline at end of file -- Gitee