Open bingjiw opened 8 months ago
Can you run the example videos on 40GB on Colab? My example video requires less than 24 GB on my 4090 GPU.
Yes, the example of "run_fresco.ipynb" is sucessfully running on Colab with A100 with 40GB GPU RAM.
What is the resolution of the video that causes your OOM?
1280x720.mp4 and 640x360.mov
Then it will be resized to 512x896 Compared to 512*512 (example video), it has 1.75X pixels Then, in our feature optimization, we need to compute the gram matrix, which will be 1.75x1.75=3.0625X memory. Causing OOM.
Maybe you can use smaller batch size or resize video to smaller resolution. https://github.com/williamyang1991/FRESCO/blob/b3148c210f85be60fb10d4e51548c6b65f03f80b/config/config_carturn.yaml#L20 https://github.com/williamyang1991/FRESCO/blob/9fe1be71b6c21890b5bc92659026f9586440266e/run_fresco.py#L170
thank you
run webUI.py on Colab with A100 GPU and 40G GPU RAM still got Error:
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 12.25 GiB (GPU 0; 39.56 GiB total capacity; 21.83 GiB already allocated; 10.45 GiB free; 22.58 GiB 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