search

wenqsun / DimensionX

DimensionX: Create Any 3D and 4D Scenes from a Single Image with Controllable Video Diffusion
Apache License 2.0
913 stars 54 forks source link

CogVideoX1.5-5B-I2V or "THUDM/CogVideoX-5b-I2V" #19

Open keyonchrist opened 6 days ago

keyonchrist commented 6 days ago

Are these lora weights for THUDM/CogVideoX-5b-I2V or CogVideoX1.5-5B-I2V, or are they the same version? Thanks and amazing work guys! your paper is awesome

chenshuo20 commented 6 days ago

We trained these LoRAs using the THUDM/CogVideoX-5b-I2V model. Interestingly, they also appear to be compatible with CogVideoX1.5-5B-I2V. Feel free to give it a try yourself!

keyonchrist commented 5 days ago

ok will do!

keyonchrist commented 4 days ago

We trained these LoRAs using the THUDM/CogVideoX-5b-I2V model. Interestingly, they also appear to be compatible with CogVideoX1.5-5B-I2V. Feel free to give it a try yourself!

Hey! I tried to run the model but keep getting this error, any idea why?

" File "/home/ubuntu/venv/lib/python3.10/site-packages/accelerate/utils/modeling.py", line 247, in set_module_tensor_to_device new_module = getattr(module, split) File "/home/ubuntu/venv/lib/python3.10/site-packages/diffusers/models/modeling_utils.py", line 155, in getattr return super().getattr(name) File "/home/ubuntu/venv/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1931, in getattr raise AttributeError( AttributeError: 'CogVideoXTransformer3DModel' object has no attribute 'ofs_embedding'. Did you mean: 'time_embedding'?"

specifically the last line: AttributeError: 'CogVideoXTransformer3DModel' object has no attribute 'ofs_embedding'. Did you mean: 'time_embedding'?"

chenshuo20 commented 4 days ago

It appears that the issue stems from CogVideoX1.5-5B-I2V, an additional component compared to CogVideoX-5B-I2V. You might resolve this by upgrading your diffusers version, trying cloning their latest repository and installing it. I also find a similar issue reported in the CogVideo repo.