Ive tried running the fune tuning with multi_temporal_crop_classification.py on both V100 and A10G but get CUDA out of memory. Can you advise if a machine with larger memory is required, or if there is a way to modify the training config but still reproduce the paper conditions?
Thanks
Ive tried running the fune tuning with
multi_temporal_crop_classification.py
on both V100 and A10G but getCUDA out of memory
. Can you advise if a machine with larger memory is required, or if there is a way to modify the training config but still reproduce the paper conditions? Thanks