TonariNoTaku commited on
Commit
a0e0090
·
1 Parent(s): fe87140

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +5 -0
README.md CHANGED
@@ -9,6 +9,7 @@ tags:
9
  This model is a LORA for improving humanoid women bodies in SDXL.
10
 
11
  Dattaset:
 
12
  It was learned on ~500 perfect pictures (without any unperfections) with repetition 10 * batch 5 * 4...8 epoh, and ~500...1500 very good pictures (with most features very good) for diversity with lower attention (repetition 1..4).
13
  All womens in range 18-26, most was 20yo.
14
 
@@ -26,7 +27,11 @@ Failed attempts:
26
  * LoKR recommended Multires noise discount 0.8 is not working from start
27
 
28
  Version history:
 
29
  all standart LoRA, prodigy preset, scale buckets, no regularization images
 
30
  v0.25 no spec activation. Partially fixed problematic features. 1750 images, network 128/96, scale_weight_norms=2
 
31
  v0.2 no spec activation. may or not use 1girl. 1300 images. learned with prodigy and adafactor preset, network 128/16. Same problematic lips and nipples due dataset diversity and probably overtraining on some pictures.
 
32
  v0.1 activation: g01tt. 1100 images, network 32/16, but has some overtraining issues but undertrained in same time. Problematic nipples.
 
9
  This model is a LORA for improving humanoid women bodies in SDXL.
10
 
11
  Dattaset:
12
+
13
  It was learned on ~500 perfect pictures (without any unperfections) with repetition 10 * batch 5 * 4...8 epoh, and ~500...1500 very good pictures (with most features very good) for diversity with lower attention (repetition 1..4).
14
  All womens in range 18-26, most was 20yo.
15
 
 
27
  * LoKR recommended Multires noise discount 0.8 is not working from start
28
 
29
  Version history:
30
+
31
  all standart LoRA, prodigy preset, scale buckets, no regularization images
32
+
33
  v0.25 no spec activation. Partially fixed problematic features. 1750 images, network 128/96, scale_weight_norms=2
34
+
35
  v0.2 no spec activation. may or not use 1girl. 1300 images. learned with prodigy and adafactor preset, network 128/16. Same problematic lips and nipples due dataset diversity and probably overtraining on some pictures.
36
+
37
  v0.1 activation: g01tt. 1100 images, network 32/16, but has some overtraining issues but undertrained in same time. Problematic nipples.