torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.56 GiB (GPU 0; 23.68 GiB total capacity; 18.96 GiB already allocated; 2.84 GiB free; 18.97 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.

Dear Authors, Thanks a lot for the awesome work on CogVideo! I was wondering do you plan to release the model weights fine-tuned on UCF or kinetics? Thanks a lot!

My GPU is RTX3080, but when I use the command sudo sh ./scripts/inference_cogvideo_pipeline.sh, the following error occurs RuntimeError: CUDA out of memory. Tried to allocate 54.00 MiB (GPU 0; 9.78 GiB total capacity; 9.53 GiB already allocated; 28.31 MiB free; 9.54 GiB reserved in total by PyTorch)...

Can you tell me how to solve it? Thanks.

all log in https://colab.research.google.com/drive/13oB5IP54VzlNM9Tzyuv9Un_vOBZNWT77 env : GPU Tesla T4 15360MiB cuda 11.6 torch 1.13.1+cu116 already update --batch-size 1 --max-inference-batch-size 1 \

I did't know why

Hello! I am using a p3.8xlarge ec2 instance with GPU mem of 64gb (https://aws.amazon.com/ec2/instance-types/p3/) I have reduced batch-size and max-inference-batch-size to 1 and also installed: pip install git+https://github.com/Sleepychord/Image-Local-Attention however I am still getting this OOM, w...

environment: torch version： 1.13.0+cu117 cuda: 11.6 detail info: _INFO:root:[Generating First Frame with CogView2]Raw text: 一个男人在滑雪 高清摄影 Traceback (most recent call last): File "cogvideo_pipeline.py", line 793, in main(args) File "cogvideo_pipeline.py", line 736, in main parent_given_to...

So the old CogVideo space on Hugging Face was removed. I'm not using Replicate because I need a credit card to continue. Please, make a new demo on CogVideo.

Hi, when I run scripts 'sh scripts/inference_cogvideo_pipeline.sh' Error occurs