jim60105 commited on
Commit
23dd609
·
verified ·
1 Parent(s): 394ec04

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +6 -9
README.md CHANGED
@@ -50,20 +50,18 @@ license: cc-by-nc-nd-4.0
50
 
51
  This character is the Vtuber 薬袋アルマ from GuildCQ [https://www.youtube.com/@ArumaCh](https://www.youtube.com/@ArumaCh)
52
 
53
- Trained on Kohaku-XL Delta and Animagine XL V3.1 with 20 images.
54
-
55
- The base trigger word is MinaiAruma, 1girl
56
-
57
- Use `waistcoat` / `strapless` to control her blue vest(?) on the top half of her body.
58
 
 
 
59
  Use `detached sleeves` / `sleeveless` to control the sleeves.
60
 
61
  ---
62
 
63
- Please note that this model is slightly overtrained and it may cause failures above LoRA strength 1.2. When its concept is not strong enough, try adding weight to the trigger word MinaiAruma first rather than the LoRA. In other words:
64
-
65
- Try this: **(MinaiAruma:1.3), 1girl, <lora:MinaiArumaXL_kohaku:1>**
66
 
 
 
67
  Not this: **(MinaiAruma:1), 1girl, <lora:MinaiArumaXL_kohaku:1.3>**
68
 
69
  ---
@@ -74,7 +72,6 @@ Adhering to [GuildCQ Terms of Use](https://twpf.jp/GCQ_Official), please do not
74
 
75
  You should use `MinaiAruma, 1girl` to trigger the image generation.
76
 
77
-
78
  ## Download model
79
 
80
  Weights for this model are available in Safetensors format.
 
50
 
51
  This character is the Vtuber 薬袋アルマ from GuildCQ [https://www.youtube.com/@ArumaCh](https://www.youtube.com/@ArumaCh)
52
 
53
+ Trained on [Kohaku-XL Delta](https://huggingface.co/KBlueLeaf/Kohaku-XL-Delta) and [Animagine XL V3.1](https://huggingface.co/cagliostrolab/animagine-xl-3.1) with 20 images.
 
 
 
 
54
 
55
+ The base trigger word is `MinaiAruma, 1girl`
56
+ Use `waistcoat` / `strapless` to control her blue vest(?) on the top half of her body.
57
  Use `detached sleeves` / `sleeveless` to control the sleeves.
58
 
59
  ---
60
 
61
+ Please note that this model is slightly overtrained and it may cause failures above LoRA strength 1.2. When its concept is not strong enough, try adding weight to the trigger word MinaiAruma first rather than the LoRA.
 
 
62
 
63
+ In other words:
64
+ Try this: **(MinaiAruma:1.3), 1girl, <lora:MinaiArumaXL_kohaku:1>**
65
  Not this: **(MinaiAruma:1), 1girl, <lora:MinaiArumaXL_kohaku:1.3>**
66
 
67
  ---
 
72
 
73
  You should use `MinaiAruma, 1girl` to trigger the image generation.
74
 
 
75
  ## Download model
76
 
77
  Weights for this model are available in Safetensors format.