UnifiedHorusRA commited on
Commit
bbcd857
·
verified ·
1 Parent(s): a6c9564

Upload v2.0_Qwen/README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. v2.0_Qwen/README.md +73 -0
v2.0_Qwen/README.md ADDED
@@ -0,0 +1,73 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ tags:
5
+ - art
6
+ ---
7
+
8
+ # A Better Wolf
9
+
10
+ **Creator**: [Torque](https://civitai.com/user/Torque)
11
+ **Type**: LORA
12
+ **Base Model**: Qwen
13
+ **Version**: v2.0 Qwen
14
+ **Trigger Words**: `N/A`
15
+
16
+ **Civitai Model ID**: 412694
17
+ **Civitai Version ID**: 2123452
18
+
19
+ **Stats (at time of fetch for this version)**:
20
+ * Downloads: 104
21
+ * Rating: 0 (0 ratings)
22
+ * Favorites: N/A
23
+
24
+ ---
25
+
26
+ ## 📄 Description (Parent Model)
27
+
28
+ Have you noticed that when a wolf snarls you can see its gums but most AI models don't show this?
29
+ Trained on 426 photos of real wolves (no art, no werewolves nor AI gen), tagged manually and using GPT4 to 75 tokens each.
30
+ The main trigger word is
31
+ wolf
32
+ so Just add this lora and prompt as usual and get - A Better Wolf™ -
33
+ Qwen
34
+ I've enhanced the dataset to a v2 using 689 images and joycaption to meticulously describe each scene. then trained I trained using Ostris new
35
+ 3 bit
36
+ quant and side chain "accuracy recovery adapter" . The LoRA is
37
+ CLIP-heavy
38
+ : it defines shape and features, not style. so load it onto the CLIP side of your workflow. i use a modified workflow from Aitrepreneur and can't really share it.
39
+ Flux:
40
+ We have 3 versions two of which works and the third is for creating nightmares.
41
+ Recommended
42
+ is trained on 750 steps and gives a very impressive result
43
+ Light
44
+ is lighter trained at 500 steps and can add detail to a scene
45
+ - I've kept all test images with seed = 1 and prompts in same order so you can compare the two models -
46
+ Bad Taxidermy
47
+ can create good results, which is why I uploaded it but it turned out to be very over trained and give absolutely horrible images. so don't download it unless you need that.
48
+ Trained using
49
+ ostris ai-toolkit
50
+ SDXL:
51
+ This LORA has been tested quite a lot now and it fixes a lot of the wolf related issues in SDXL
52
+ NB Training data included
53
+
54
+ ## Version Notes (v2.0 Qwen)
55
+
56
+ Qwen version trained on my V2 dataset.. which has a lot better tagging
57
+
58
+ ---
59
+
60
+ ## Civitai Links
61
+
62
+ * **[🔗 View This Version on Civitai →](https://civitai.com/models/412694?modelVersionId=2123452)**
63
+ * [View Full Model Page →](https://civitai.com/models/412694)
64
+ * [View Creator Profile →](https://civitai.com/user/Torque)
65
+
66
+ ---
67
+
68
+ ## File Information
69
+
70
+ * **Filename**: `betterwolf_000002750.safetensors`
71
+ * **Size**: 281.47 MB
72
+ * **Hash (AutoV2)**: `FCBDB5E505`
73
+ * **Hash (SHA256)**: `FCBDB5E50505F4D253BEBEEEB9F3D5380CE51E904A559FD3DAF21F257BE60747`