Hello, sorry to disturb you again,, I would like to ask how much memory your training is running under. My gpu is rtx3070, but after setting batchsize to 1, there will still be excess memory （CUDA out of memory. Tried to allocate 120.00 MiB (GPU 0; 7.76 GiB total capacity; 5.30 GiB already allocated; 67.94 MiB free; 5.56 GiB reserved in total by PyTorch)）
Hello, sorry to disturb you again,, I would like to ask how much memory your training is running under. My gpu is rtx3070, but after setting batchsize to 1, there will still be excess memory （CUDA out of memory. Tried to allocate 120.00 MiB (GPU 0; 7.76 GiB total capacity; 5.30 GiB already allocated; 67.94 MiB free; 5.56 GiB reserved in total by PyTorch)）