open-mmlab / PIA

[CVPR 2024] PIA, your Personalized Image Animator. Animate your images by text prompt, combing with Dreambooth, achieving stunning videos. PIA,你的个性化图像动画生成器,利用文本提示将图像变为奇妙的动画
https://pi-animator.github.io/
Apache License 2.0
808 stars 67 forks source link

Please create demo colab #4

Closed G-force78 closed 6 months ago

G-force78 commented 6 months ago

I am unable to get it to work based off the install instructions

Traceback (most recent call last): File "/content/PIA_env/PIA/inference.py", line 10, in from animatediff.pipelines import I2VPipeline File "/content/PIA_env/PIA/animatediff/pipelines/init.py", line 1, in from .i2v_pipeline import I2VPipeline File "/content/PIA_env/PIA/animatediff/pipelines/i2v_pipeline.py", line 10, in from diffusers.loaders import IPAdapterMixin, TextualInversionLoaderMixin ImportError: cannot import name 'IPAdapterMixin' from 'diffusers.loaders' (/usr/local/lib/python3.10/dist-packages/diffusers/loaders.py)

LeoXing1996 commented 6 months ago

Hey @G-force78. You need install diffusers==0.24.0 to run PIA. You can check your diffusers version by pip list | grep diffusers or conda list diffusers.

darkzbaron commented 6 months ago

+1

G-force78 commented 6 months ago

Installed diffusers then got error 'no module torch._dynamo'

LeoXing1996 commented 6 months ago

Hey @G-force78 and @darkzbaron, I found a third-party colab implementation: https://github.com/camenduru/PIA-colab, you may have a try.