Ziyuan Wang

Ask for help: How to use multi GPUs? All input tensors must be on the same device. Received cuda:3 and cuda:2

4

I tried to use multi gpus with data_parallel, but got an error: File "main_transformer_classification.py", line 45, in train out, att_map = self.model(data) File "/usr/local/lib64/python3.6/site-packages/torch/nn/modules/module.py", line 722, in _call_impl result =...

output images scaled to [-1,1] or [0,1] ?

2

(in your code of class PairedDataset) # output images scaled to -1,1 output_t = self.T(output_img) output_t = F.to_tensor(output_t) output_t = F.normalize(output_t, mean=[0.5], std=[0.5])