Open duanyuanyi opened 1 year ago
The issue has been resolved by downgrading the PyTorch version to 1.4.0. The reason was that the higher version was causing compatibility problems. By matching the corresponding versions, PyTorch and torchvision can be made compatible, ensuring smooth execution of the code.
(I2l-meshnet) [duanyuanyi@gpu demo]$ python demo.py --gpu 3 --stage param --test_epoch 8