lix19937 / dnn

Network interpretation & tensorrt deploy, yolox, detr3d, Bevformer, ViT, CRNN, GAN
5 stars 0 forks source link

error in ms_deformable_im2col_cuda: "no kernel image is available for execution on the device" #10

Open lix19937 opened 5 months ago

lix19937 commented 5 months ago

the error message the CUDA kernel was not built for your GPU architecture (sm_86) so you might need to rebuild it for this arch or install the right binary.

so need to rebuild it for this arch or install the right binary.

lix19937 commented 5 months ago

类似问题 运行pytorch 出现
CUDA error: no kernel image is available for execution on the device

GeForce RTX 3080 with CUDA capability sm_86 is not compatible with the current PyTorch installation. The current PyTorch install supports CUDA capabilities sm_37 sm_50 sm_60 sm_70. If you want to use the GeForce RTX 3080 GPU with PyTorch, please check the instructions at https://pytorch.org/get-started/locally/ RuntimeError: CUDA error: no kernel image is available for execution on the device

表明当前的PyTorch安装不支持你的GPU的CUDA计算能力
注意 通过 pip install torch 是直接下载whl包,这里torch已经将cuda 库绑定了,而不是依赖外部系统的cuda.

pip install torch==1.9.0+cu111 torchvision==0.10.0+cu111 torchaudio==0.9.0 -f https://download.pytorch.org/whl/torch_stable.html

验证

import torch
print(torch.cuda.is_available())  # 应该返回 True
print(torch.cuda.device_count())  # 应该返回你的 GPU 数量