Yuxuan Zhang

I don't think it's because of the balanced reason. I want to know if only one GPU was running the whole time (until the OOM error occurred), rather than multiple...

This code is completely correct, but it seems that it cannot run on multiple AMDGPUs pipe.enable_model_cpu_offload() pipe.enable_sequential_cpu_offload() Can these two sentences enable you to run on an single AMD GPU?...

No, the 3.3GB refers to the data obtained from testing on NVIDIA GPUs after using pipe.enable_sequential_cpu_offload(). In the data explanation, we also mentioned that the tests were conducted on A100/H100...

Yes, it is mutually exclusive, only one can be chosen, 64GB of free RAM is sufficient, the issue does not lie here

For multi-GPU cards, due to the design in the diffuser library, pipe.enable_sequential_cpu_offload() must be disabled. This is not something that can be determined by the model layer and requires adherence...

说的是微调吗，推理早就通了

已经在PR了，最后等待合并 https://github.com/huggingface/diffusers/pull/9302

this exports the entire model because torch.save was used when saving.

check here https://github.com/THUDM/CogVideo/tree/main/sat#using-the-fine-tuned-model

Not at the moment. Currently, we are all working on video generation models that are under 10B in size.