Model Card for maskgit-large-imagenet
<!-- Provide a quick summary of what the model is/does. -->
A MaskGit model trained on imagenet as part of the open-muse effort.
Model Details
Model Description
A maskgit-large transformer model with the following configuration:
| num_layers | num_heads | hidden_size | intermediate_size |
|---|---|---|---|
| 24 | 16 | 768 | 3072 |
The model is trained on the imagenet dataset for class conditional image generation.
Model Sources
<!-- Provide the basic links for the model. -->
- Repository: open-muse
- Paper: MaskGIT: Masked Generative Image Transformer
How to Get Started with the Model
First clone and install the open-muse repo per the readme's instructions. And use the following code snippet:
from muse import PipelineMuse
pipe = PipelineMuse.from_pretrained("openMUSE/maskgit-large-imagenet", is_class_conditioned=True).to("cuda")
class_ids = 47
images = pipe(class_ids=class_ids, num_images_per_prompt=4, timesteps=8, use_maskgit_generate=True)
Some non-cherry-picked samples generated by the model:
For more samples, check out this page.
Training Details
Training Data
The model is trained on the imagenet dataset using the class ids as conditioning.
Training Procedure
Please refer to the repo and wandb logs for details about the training.
Evaluation
TODO: