Update README.md
Browse files
README.md
CHANGED
|
@@ -4,6 +4,8 @@ license: creativeml-openrail-m
|
|
| 4 |
|
| 5 |
https://huggingface.co/spaces/CompVis/stable-diffusion-license
|
| 6 |
|
|
|
|
|
|
|
| 7 |
# New v5.1 model
|
| 8 |
|
| 9 |
The new version is trained from a basis of the RunwayML 1.5 ckpt. This fine tuning sheds the last remnant of the concepts in original DreamBooth paper as regularization via generated images is dropped in favor of a mix a scrape of laion to protect the model's original qualities instead. 1636 training images, 1636 ground truth images from laion were trained for 19009 steps at LR 4e-7.
|
|
@@ -44,8 +46,6 @@ This is a step away from textual inversion and dreambooth and towards full fine
|
|
| 44 |
|
| 45 |
Supplementary information here: https://gist.github.com/victorchall/67bc53472f86641aef1ebee1e154f5d1
|
| 46 |
|
| 47 |
-
Discord: https://discord.gg/nrvPgh94cC
|
| 48 |
-
|
| 49 |
My "group" regularization set here: https://github.com/victorchall/dreambooth-group-regularization
|
| 50 |
|
| 51 |
The group set is meant to mirror the sort of content in the group training set, and was generated with a few different captions.
|
|
|
|
| 4 |
|
| 5 |
https://huggingface.co/spaces/CompVis/stable-diffusion-license
|
| 6 |
|
| 7 |
+
EveryDream Discord: https://discord.gg/nrvPgh94cC
|
| 8 |
+
|
| 9 |
# New v5.1 model
|
| 10 |
|
| 11 |
The new version is trained from a basis of the RunwayML 1.5 ckpt. This fine tuning sheds the last remnant of the concepts in original DreamBooth paper as regularization via generated images is dropped in favor of a mix a scrape of laion to protect the model's original qualities instead. 1636 training images, 1636 ground truth images from laion were trained for 19009 steps at LR 4e-7.
|
|
|
|
| 46 |
|
| 47 |
Supplementary information here: https://gist.github.com/victorchall/67bc53472f86641aef1ebee1e154f5d1
|
| 48 |
|
|
|
|
|
|
|
| 49 |
My "group" regularization set here: https://github.com/victorchall/dreambooth-group-regularization
|
| 50 |
|
| 51 |
The group set is meant to mirror the sort of content in the group training set, and was generated with a few different captions.
|