search

gpt-omni / mini-omni

open-source multimodal large language model that can hear, talk while thinking. Featuring real-time end-to-end speech input and streaming audio output conversational capabilities.
https://arxiv.org/abs/2408.16725
MIT License
3.04k stars 269 forks source link

win10 部署后提示这个错误 AssertionError: Torch not compiled with CUDA enabled #35

Closed xie236393 closed 1 month ago

xie236393 commented 1 month ago

运行此命令之后 python server.py --ip '0.0.0.0' --port 60808

Traceback (most recent call last): File "I:\mini-omni\inference.py", line 666, in test_infer() File "I:\mini-omni\inference.py", line 516, in test_infer fabric, model, text_tokenizer, snacmodel, whispermodel = load_model(ckpt_dir, device) File "I:\mini-omni\inference.py", line 350, in load_model snacmodel = SNAC.from_pretrained("hubertsiuzdak/snac_24khz").eval().to(device) File "i:\mini-omni\venv\lib\site-packages\torch\nn\modules\module.py", line 1173, in to return self._apply(convert) File "i:\mini-omni\venv\lib\site-packages\torch\nn\modules\module.py", line 779, in _apply module._apply(fn) File "i:\mini-omni\venv\lib\site-packages\torch\nn\modules\module.py", line 779, in _apply module._apply(fn) File "i:\mini-omni\venv\lib\site-packages\torch\nn\modules\module.py", line 779, in _apply module._apply(fn) [Previous line repeated 2 more times] File "i:\mini-omni\venv\lib\site-packages\torch\nn\modules\module.py", line 804, in _apply param_applied = fn(param) File "i:\mini-omni\venv\lib\site-packages\torch\nn\modules\module.py", line 1159, in convert return t.to( File "i:\mini-omni\venv\lib\site-packages\torch\cuda__init__.py", line 284, in _lazy_init raise AssertionError("Torch not compiled with CUDA enabled") AssertionError: Torch not compiled with CUDA enabled win10 3090 依赖以按照readme安装完毕。 请问这个怎么解决呢？

BrenchCC commented 1 month ago

没安装CUDA版的Pytorch，用这段代码测试一下：

import torch
print(torch.cuda.is_available())
print(torch.cuda.get_device_name(0))
xie236393 commented 1 month ago

好滴 谢谢

pli2014 commented 1 month ago

pip install torch==2.2.1+cu121 torchaudio==2.2.1+cu121 torchvision==0.17.1+cu121 -f https://mirror.sjtu.edu.cn/pytorch-wheels/torch_stable.html -i https://pypi.tuna.tsinghua.edu.cn/simple/ some-package --trusted-host mirrors.aliyun.com

you need to a right version for torch, torchaudio, torchvsion.

kunci115 commented 1 month ago

use this fork branch for windows, tested well end to end no issue https://github.com/gpt-omni/mini-omni/pull/44 image

mini-omni commented 1 month ago

I will close it for now, please feel free to re-open.