PyTorch CUDA Conda/Pip Init

Table

版本

https://download.pytorch.org/whl/torch_stable.html

	torch	torchvision
cu90	0.3.0, 0.3.1, 0.4.0, 0.4.1, 1.0.[01], 1.1.1
cu91	0.3.1, 0.4.0
cu92	0.4.1, 1.2.0, 1.3.[01], 1.4.0, 1.5.[01], 1.6.0, 1.7.[01]
cu100	1.3.0, 1.4.0, 1.5.0	0.4.1, 0.4.2, 0.4.3 0.5.0
cu101	1.3.0, 1.3.1, 1.4.0, 1.5.0, 1.5.1 1.6.0, 1.7.0, 1.7.1, 1.8.0, 1.8.1
cu102	1.5.0–1.12.1
cu110	1.7.0,1.7.1
cu111	1.8.[01], 1.9.[01] ,1.10.[012]
cu113	1.10, 1.11, 1.12,
cu115	1.11
cu116	1.12, 1.13,
cu117	1.13, 2.0.0, 2.0.1
cu118	2.0.0, 2.0.1, 2.1.0, 2.1.1, 2.1.2, 2.2.0, 2.2.1
cu121	2.1.[012], 2.2.[01]

LibTorch

Conda

pytorch	CUDA
1.5.1_cu92	conda install pytorch==1.5.1 torchvision==0.6.1 cudatoolkit=9.2 -c pytorch
1.5.1_cu101	conda install pytorch==1.5.1 torchvision==0.6.1 cudatoolkit=10.1 -c pytorch
1.5.1_CPU	conda install pytorch==1.5.1 torchvision==0.6.1 cpuonly -c pytorch
1.10.1_cu102	conda install pytorch==1.10.1 torchvision==0.11.2 torchaudio==0.10.1 cudatoolkit=10.2 -c pytorch
1.8.0_cu111_docker	docker pull pytorch/pytorch:1.8.0-cuda11.1-cudnn8-devel
1.8.0_cu111	conda install pytorch==1.8.0 torchvision==0.9.0 torchaudio==0.8.0 cudatoolkit=11.1 -c pytorch -c conda-forge	-
2.1.2_cu118	conda install pytorch==2.1.2 torchvision==0.16.2 torchaudio==2.1.2 cudatoolkit=11.8 -c pytorch -c nvidia
	conda install pytorch==2.1.2 torchvision==0.16.2 torchaudio==2.1.2 pytorch-cuda=11.8 -c pytorch -c nvidia

pytorch-lightning

安装pytorch-lightning时一定注意自己的torch是pip安装还是conda安装，pytorch_lightning 安装方式要与torch的安装方式保持一致，否则也会导致你的torch版本被替换。

conda install pytorch-lightning -c conda-forge

Pip

CUDA 11.8	pip3 install torch torchvision torchaudio –index-url https://download.pytorch.org/whl/cu118
CPU	pip3 install torch torchvision torchaudio –index-url https://download.pytorch.org/whl/cpu
CUDA-92	https://download.pytorch.org/whl/cu92

版本关系

	pytorch	xformers	differusers	transformers
	torch==2.2.2	0.0.25.post1
	torch==2.2.1	0.0.25		4.36.2