Closed hujun1399 closed 1 year ago
hello! I am trying to reproduce the 288 model by setting image_size to 288*288 and batch_size to 16. When I run wav2lip_train.py, the GPU memory usage reaches over 30GB. Is this normal? And is there any way to reduce the memory usage?
My GPU is a 3070 (with a memory upgrade to 24GB). After changing the batch size to 8, it can start training.
smaller batch size should work
hello! I am trying to reproduce the 288 model by setting image_size to 288*288 and batch_size to 16. When I run wav2lip_train.py, the GPU memory usage reaches over 30GB. Is this normal? And is there any way to reduce the memory usage?