Downloads · 30 days
3
1% of all-time downloads
ddpo-incompressibility
kvablack/ddpo-incompressibility
ddpo-incompressibility is a text-to-image model from kvablack. Use it when you need an image from a text prompt. It is set up for diffusers. The card lists the license as creativeml-openrail-m.
This model was finetuned from Stable Diffusion v1-4 using DDPO and a reward function encouraging images that are not JPEG-compressible. See the project website for more details.
All-time downloads
270
Public
Repo size
5.5 GB
Likes
1
Public
Repo makeup
Click a slice to open those files.
.bin5.5 GB · 100%
At a glance
- Task
- Text-to-Image
- Library
- diffusers
- License
- creativeml-openrail-m
- Access
- Public
- Created
- May 26, 2023
- Updated
- May 30, 2023
- SHA
- 50b5e007
Model card
From the Hugging Face model README
ddpo-incompressibility
This model was finetuned from Stable Diffusion v1-4 using DDPO and a reward function encouraging images that are not JPEG-compressible. See the project website for more details.
The model was finetuned for 20 iterations with a batch size of 256 samples per iteration. During finetuning, it was prompted with all of the animals in the Imagenet-1000 categories (the first 398 categories), but it exhibits some generalization to other prompts.
- Task
- Text-to-Image
- Library
- diffusers
- License
- creativeml-openrail-m
- Languages
- en
- Created
- May 26, 2023
- Updated
- May 30, 2023