Train simultaneously on two datasets

43

Did you forget to add the if-clause protection as explained in the error message?
If so, did its usage fix the error?
Here is a small example:

import torch

def main()
    for i, data in enumerate(dataloader):
        # do something here

if __name__ == '__main__':
    main()
44

yes, i have added it also…now im getting different OS Error
OSError: [WinError 1455] The paging file is too small for this operation to complete. Error loading "C:\Users\test\Anaconda3\envs\py36\lib\site-packages\torch\lib\caffe2_detectron_ops_gpu.dll" or one of its dependencies.

45 
class ConcatDataset(Dataset):
    def __init__(self,dataset_list, probs=None):
        self.datasets = dataset_list
        self.probs = probs if probs else [1/len(self.datasets)] * len(self.datasets)

    def __getitem__(self, i):
        chosen_dataset = np.random.choice(self.datasets, 1, p=self.probs)[0]
        i = i % len(chosen_dataset)
        return chosen_dataset[i]

    def __len__(self):
        return  max(len(d) for d in self.datasets)
46

Maybe you can try to set the random seed before each iteration

def setup_seed(seed):
    torch.manual_seed(seed)
    torch.cuda.manual_seed_all(seed)
    np.random.seed(seed)
    random.seed(seed)
    torch.backends.cudnn.deterministic = True

setup_seed(123)
inputs1, target1 = next(iter(train_dl1))
setup_seed(123)
inputs2, targets2 = next(iter(train_dl2))