error will executing /inference.py Error PYTORCH_CUDA_ALLOC_CONF

[jeevitesh]

Hi ,
Please help me to resolve the issue moment i run the program i get this error
new_node1 = torch.matmul(res_feature_after_view1, self.node_fea_for_res)
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.41 GiB (GPU 0; 2.00 GiB total capacity; 330.44 MiB already allocated; 1.23 GiB free; 410.00 MiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

approach tried

1. import os
   os.environ["PYTORCH_CUDA_ALLOC_CONF"] = "max_split_size_mb:2000" -> did'nt work
2. os.environ["PYTORCH_CUDA_ALLOC_CONF"] = 'max_split_size_mb:2000' -> did'nt work
3. torch.cuda.empty_cache() -> in the beginning of the file and several places within the loop
   for iii, sample_batched in enumerate(zip(testloader_list, testloader_flip_list)):