Zishen Huang

> > My CUDA version is 12.1. > > 1.pip install cupy-cuda12x. 2.conda install pytorch==2.0.1 torchvision==0.15.2 torchaudio==2.0.2 pytorch-cuda=11.8 -c pytorch -c nvidia. > > To modify compile_with_cache function: > >...

> > > My CUDA version is 12.1. > > > 1.pip install cupy-cuda12x. 2.conda install pytorch==2.0.1 torchvision==0.15.2 torchaudio==2.0.2 pytorch-cuda=11.8 -c pytorch -c nvidia. > > > To modify compile_with_cache...

> > > > My CUDA version is 12.1. > > > > 1.pip install cupy-cuda12x. 2.conda install pytorch==2.0.1 torchvision==0.15.2 torchaudio==2.0.2 pytorch-cuda=11.8 -c pytorch -c nvidia. > > > >...

第85行，`pipe.fuse_lora(lora_scale=1 / lora_rank)`改成`pipe.fuse_lora(lora_scale=1 , components=['transformer'])`试试，我是这样跑出效果的

> > 第85行，`pipe.fuse_lora(lora_scale=1 / lora_rank)`改成`pipe.fuse_lora(lora_scale=1 , components=['transformer'])`试试，我是这样跑出效果的 > > lora注入到视频生成，控制什么呢？效果怎样 控制人物形象，5b效果挺好的，然后风格训练不了，要6000个视频，内存2t直接爆了，人物200个视频就可以训练好了 https://github.com/user-attachments/assets/86d46389-c123-4e64-8976-cff1d4e2ceca 下面是参考人物 https://github.com/user-attachments/assets/a889f39b-1711-4c23-bd1d-0a40c39afbcf

> We have update [requirements.txt](https://github.com/PKU-YuanGroup/ConsisID/blob/main/requirements.txt), please reinstall the env. Or you can try `pip install git+https://github.com/SHYuanBest/ConsisID_diffusers.git`. It seems that your work already supports diffusers. May I ask if the model...