Windows10下PyTorch无CUDA编译支持,GPU无法用于ML任务求助
PyTorch CUDA不可用问题解决(Windows 10 x64)
问题场景
运行以下代码尝试用GPU处理机器学习任务时触发错误:
import torch from transformers import AutoTokenizer, AutoModel tokenizer = AutoTokenizer.from_pretrained("cointegrated/rubert-tiny2") model = AutoModel.from_pretrained("cointegrated/rubert-tiny2") model.cuda() # uncomment it if you have a GPU def embed_bert_cls(text, model, tokenizer): t = tokenizer(text, padding=True, truncation=True, return_tensors='pt') with torch.no_grad(): model_output = model(**{k: v.to(model.device) for k, v in t.items()}) embeddings = model_output.last_hidden_state[:, 0, :] embeddings = torch.nn.functional.normalize(embeddings) return embeddings[0].cpu().numpy()
错误信息:
AssertionError: Torch not compiled with CUDA enabled
环境检查结果
torch.cuda.is_available()返回falsetorch.__version__返回2.0.1+cpunvcc --version输出:
nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2023 NVIDIA Corporation Built on Mon_Apr__3_17:36:15_Pacific_Daylight_Time_2023 Cuda compilation tools, release 12.1, V12.1.105 Build cuda_12.1.r12.1/compiler.32688072_0
nvidia-smi输出:
+---------------------------------------------------------------------------------------+ | NVIDIA-SMI 531.68 Driver Version: 531.68 CUDA Version: 12.1 | |-----------------------------------------+----------------------+----------------------+
已尝试操作
卸载所有pip库后执行以下命令重装,但PyTorch仍为CPU版本:
pip install torch torchvision torchaudio -f https://download.pytorch.org/whl/cu121/torch_stable.html
解决步骤(仅用pip)
彻底清理PyTorch残留
- 卸载现有组件:
pip uninstall -y torch torchvision torchaudio - 手动删除Python环境
Lib/site-packages目录下的torch、torchvision、torchaudio文件夹(若存在)
- 卸载现有组件:
指定CUDA版本安装GPU版PyTorch
针对CUDA 12.1,执行以下命令安装适配版本:pip install torch==2.0.1+cu121 torchvision==0.15.2+cu121 torchaudio==2.0.2+cu121 --index-url https://download.pytorch.org/whl/cu121该命令明确指定了CUDA 12.1对应的GPU兼容包,避免默认安装CPU版。
验证安装
运行以下代码确认:import torch print(torch.__version__) # 预期输出:2.0.1+cu121 print(torch.cuda.is_available()) # 预期输出:True
内容的提问来源于stack exchange,提问作者Evgeniy Shornikov
相关产品推荐
相关产品推荐

