PaddlePaddle / PaddleSpeech

Easy-to-use Speech Toolkit including Self-Supervised Learning model, SOTA/Streaming ASR with punctuation, Streaming TTS with text frontend, Speaker Verification System, End-to-End Speech Translation and Keyword Spotting. Won NAACL2022 Best Demo Award.
https://paddlespeech.readthedocs.io
Apache License 2.0
10.92k stars 1.83k forks source link

安装paddlespeech报错 #2521

Open will-wiki opened 1 year ago

will-wiki commented 1 year ago

安装方法 安装指导中的【简单】方法, paddlepaddle-gpu==2.3.0版本 安装命令如下: pip install pytest-runner -i https://pypi.tuna.tsinghua.edu.cn/simple pip install paddlespeech -i https://pypi.tuna.tsinghua.edu.cn/simple

出现报错信息: ERROR: Cannot install paddlespeech==0.0.1a0, paddlespeech==0.0.2a0, paddlespeech==0.1.0a0, paddlespeech==0.1.0a1, paddlespeech==0.1.0a2, paddlespeech==0.1.0a3, paddlespeech==0.1.0a4, paddlespeech==0.1.0a5 and paddlespeech==0.1.0a6 because these package versions have conflicting dependencies.

The conflict is caused by: paddlespeech 0.1.0a6 depends on paddlespeech-ctcdecoders paddlespeech 0.1.0a5 depends on paddlespeech-ctcdecoders paddlespeech 0.1.0a4 depends on paddlespeech-ctcdecoders paddlespeech 0.1.0a3 depends on paddlespeech-ctcdecoders paddlespeech 0.1.0a2 depends on paddlespeech-ctcdecoders paddlespeech 0.1.0a1 depends on paddlespeech-ctcdecoders paddlespeech 0.1.0a0 depends on paddlespeech-ctcdecoders paddlespeech 0.0.2a0 depends on paddlespeech-ctcdecoders paddlespeech 0.0.1a0 depends on paddlespeech-ctcdecoders

To fix this you could try to:

  1. loosen the range of package versions you've specified
  2. remove package versions to allow pip attempt to solve the dependency conflict
yt605155624 commented 1 year ago

升级 pip, 你的 pip 没有找到最新的 paddlespeech

will-wiki commented 1 year ago

@yt605155624 用pip install --upgrade pip尝试更新了pip,好像版本是最新的21.3.1,但是还是报同样的错误,有可能是paddlepaddle=2.3.0的问题么

yt605155624 commented 1 year ago

应该不是,你安装 paddlespeech 的时候指定下版本试试 1.2.0

will-wiki commented 1 year ago

@yt605155624 好像没有你说的版本 pip install paddlespeech==1.2.0 -i https://pypi.tuna.tsinghua.edu.cn/simple 报错 Looking in indexes: https://pypi.tuna.tsinghua.edu.cn/simple ERROR: Could not find a version that satisfies the requirement paddlespeech==1.2.0 (from versions: 0.0.1a0, 0.0.2a0, 0.1.0a0, 0.1.0a1, 0.1.0a2, 0.1.0a3, 0.1.0a4, 0.1.0a5, 0.1.0a6) ERROR: No matching distribution found for paddlespeech==1.2.0

命令 pip install paddlespeech== Looking in indexes: https://mirror.baidu.com/pypi/simple ERROR: Could not find a version that satisfies the requirement paddlespeech== (from versions: 0.0.1a0, 0.0.2a0, 0.1.0a0, 0.1.0a1, 0.1.0a2, 0.1.0a3, 0.1.0a4, 0.1.0a5, 0.1.0a6) ERROR: No matching distribution found for paddlespeech==

yt605155624 commented 1 year ago

你的 pip 有问题,没有找到最新版,paddlespeech 已经更新到 1.2.0 版本了 https://pypi.org/project/paddlespeech/

liucongg commented 1 year ago

window 安装是会报错的

will-wiki commented 1 year ago

@yt605155624 直接用pip install whl包也会报错

image

看样子应该是ppython版本的问题,我这边能用的是python3.6版本,Pillow 9.0需要3.7以上,paddlespeech是一定要python3.7以上的版本才能安装么 @liucongg 实在ubuntu上安装的

will-wiki commented 1 year ago

@yt605155624 另外想问下ECAPA-TDNN模型有静态图导出脚本嘛,github里面只找到了相关的python API调用示例,看代码是每抽一个音频特征会读一遍模型?那部署起来耗时应该会比较严重吧...

yt605155624 commented 1 year ago

@will-wiki 目前没有尝试 ECAPA-TDNN 模型的动转静导出,可以参考 https://github.com/PaddlePaddle/PaddleSpeech/blob/develop/examples/csmsc/tts3/local/synthesize_e2e.sh 写下动转静代码看看能不能导出,参考文档 https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/jit/index_cn.html

python API 可以写 for 循环调用,加载一次模型调用多次,参考 vc2_infer.pyhttps://github.com/PaddlePaddle/PaddleSpeech/discussions/1256#discussioncomment-2195932

yt605155624 commented 1 year ago

@will-wiki 我们安装文档中要求了 Python >= 3.7