Arbitrary Shape Text Detection via Boundary Transformer；The paper at: https://arxiv.org/abs/2205.05320, which has been accepted by IEEE Transactions on Multimedia (T-MM 2023).
172
stars
37
forks
source link
Ctw1500训练遇到Expected a 'cuda' device type for generator but found 'cpu'解决办法 #24
报错信息：File "/home/wit/anaconda3/envs/lj_py3.8pt1.10/lib/python3.8/site-packages/torch/utils/data/sampler.py", line 126, in iter
yield from torch.randperm(n, generator=generator).tolist()
RuntimeError: Expected a 'cuda' device type for generator but found 'cpu'
原因;pytorch版本高导致，我pytorch版本为1.10
解决方案，修改文件126行代码：
yield from torch.randperm(n, generator=generator).tolist()
修改为：
yield from torch.randperm(n, generator=torch.Generator(device='cuda')).tolist()
报错信息：File "/home/wit/anaconda3/envs/lj_py3.8pt1.10/lib/python3.8/site-packages/torch/utils/data/sampler.py", line 126, in iter yield from torch.randperm(n, generator=generator).tolist() RuntimeError: Expected a 'cuda' device type for generator but found 'cpu' 原因;pytorch版本高导致，我pytorch版本为1.10 解决方案，修改文件126行代码：
yield from torch.randperm(n, generator=generator).tolist()修改为：yield from torch.randperm(n, generator=torch.Generator(device='cuda')).tolist()