Merge branch 'main' of https://huggingface.co./Aotsuyu/LoRA

.gitattributes

@@ -32,3 +32,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+diffLora/schizogrids/xyz_grid-0137-VGhhbmtzLiBJIGhhdmUgYmVlbiB0cnlpbmcgdG8gdW5kZXJzdGFuZCBnaXRodWIgZm9yIHllYXJzLg%3D%3D_2022689537.png filter=lfs diff=lfs merge=lfs -text
+diffLora/schizogrids/xyz_grid-0138-VGhhbmtzLiBJIGhhdmUgYmVlbiB0cnlpbmcgdG8gdW5kZXJzdGFuZCBnaXRodWIgZm9yIHllYXJzLg%3D%3D_4279822113.png filter=lfs diff=lfs merge=lfs -text

README.md

@@ -4,6 +4,8 @@ license: creativeml-openrail-m
 Arbitrary epochs I use the most.
 bangkisks has the same token as name, it's the outfit from 暁Records's Rock 'n' Rock 'n' Beat PV.
 
-diffLora are magical little lora trained by overfitting on an image, then editing that image and training another lora and merging the loras with -1 and 1 weights.  
+### difflora
+These are magical little lora trained by overfitting on an image, then editing that image and training another lora and merging the loras with -1 and 1 weights.  
 This achieves a small lora that doesn't affect other stuff very much, mostly just providing the effect.  
-Most also yield interesting results at negative weights, and weights exceeding 1.
+Most also yield interesting results at negative weights, and weights exceeding 1.  
+Some of them are also overfit due to how this method works (and some skill issue on my part as well, obviously), if you see a right-facing Reimu potrait, then drop the weight.