Update vae config to remove unused parameters
#9
by
a-r-r-o-w
HF staff
- opened
Hey!
These parameters are no longer supported in the Diffusers VAE implementation. They weren't used to begin with, but made its way into the codebase after reviews. They've now been remove which causes the following warning is raised when running inference with CogVideoX-2b:
The config attributes {'mid_block_add_attention': True, 'sample_size': 256} were passed to AutoencoderKLCogVideoX, but are not expected and will be ignored. Please verify your config.json configuration file.
This PR will get rid of the warning. Thanks!
zRzRzRzRzRzRzR
changed pull request status to
merged