Cognomen commited on
Commit
a214bfd
1 Parent(s): 46a9d1b

update checkpoint and readme

Browse files
Files changed (2) hide show
  1. README.md +3 -3
  2. diffusion_flax_model.msgpack +1 -1
README.md CHANGED
@@ -2,7 +2,7 @@
2
  license: mit
3
  ---
4
 
5
- Proof of concept made for the [Huggingface JAX/Diffusers community sprint](https://github.com/huggingface/community-events/tree/main/jax-controlnet-sprint)
6
 
7
  [Demo available here](https://huggingface.co/spaces/Cognomen/CatCon-Controlnet-WD-1-5-b2)
8
 
@@ -11,8 +11,8 @@ This is a controlnet for the Stable Diffusion checkpoint [Waifu Diffusion 1.5 be
11
  Major limitations:
12
 
13
  - The current checkpoint was trained on 768x768 crops without aspect ratio checkpointing. Loss in coherence for non-square aspect ratios can be expected.
14
- - The training dataset is extremely noisy and used without filtering stylistic outliers from within each category, so performance may be less than ideal.
15
  - The Waifu Diffusion base model is a hybrid anime/photography model, and can unpredictably jump between those modalities.
16
- - As styling is incredibly sensitive to divergences in model checkpoints, the capabilities of this controlnet are not expected to predictably apply to other SD 2.X checkpoints.
17
 
18
  Waifu Diffusion 1.5 beta 2 is licensed under [Fair AI Public License 1.0-SD](https://freedevproject.org/faipl-1.0-sd/). This controlnet imposes no restrictions beyond the MIT license, but it cannot be used independently of a base model.
 
2
  license: mit
3
  ---
4
 
5
+ Experimental proof of concept made for the [Huggingface JAX/Diffusers community sprint](https://github.com/huggingface/community-events/tree/main/jax-controlnet-sprint)
6
 
7
  [Demo available here](https://huggingface.co/spaces/Cognomen/CatCon-Controlnet-WD-1-5-b2)
8
 
 
11
  Major limitations:
12
 
13
  - The current checkpoint was trained on 768x768 crops without aspect ratio checkpointing. Loss in coherence for non-square aspect ratios can be expected.
14
+ - The training dataset is extremely noisy and used without filtering stylistic outliers from within each category, so performance may be less than ideal. A more diverse dataset with a larger variety of styles and categories would likely have better performance.
15
  - The Waifu Diffusion base model is a hybrid anime/photography model, and can unpredictably jump between those modalities.
16
+ - As styling is sensitive to divergences in model checkpoints, the capabilities of this controlnet are not expected to predictably apply to other SD 2.X checkpoints.
17
 
18
  Waifu Diffusion 1.5 beta 2 is licensed under [Fair AI Public License 1.0-SD](https://freedevproject.org/faipl-1.0-sd/). This controlnet imposes no restrictions beyond the MIT license, but it cannot be used independently of a base model.
diffusion_flax_model.msgpack CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:339b3dd18e3c5a477089d057e2263eeca6e121a4cd08531e34c7ac79f8cabc68
3
  size 1456925250
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:179c1b8dec7421335a16d9b9520f6d05f1f89639873a578e0158a35f232aedcb
3
  size 1456925250