SnJake commited on
Commit
1f85acb
·
verified ·
1 Parent(s): 8f341ed

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +18 -10
README.md CHANGED
@@ -15,9 +15,14 @@ tags:
15
  - safetensors
16
  ---
17
 
18
- # Ref2Font V2 — Contextual LoRA for FLUX.2 Klein 9B
19
 
20
- **Ref2Font V2** is an updated **contextual LoRA** trained for **black-forest-labs/FLUX.2-klein-9B**. It generates **1280×1280 font atlases** from a single reference image "Aa".
 
 
 
 
 
21
 
22
  **Changes in V2:**
23
  - Fixed alignment issues (letters no longer "dance" or jump).
@@ -28,12 +33,11 @@ tags:
28
 
29
  ## Examples
30
 
 
31
 
32
- ![Example_1_C](https://cdn-uploads.huggingface.co/production/uploads/658814fd586088fd274d8cc1/0r_Q2q9sNvNXKzAgUamG-.png)
33
-
34
- ![Example_2_C](https://cdn-uploads.huggingface.co/production/uploads/658814fd586088fd274d8cc1/GQbdv0phQdc8a-lZMdnI1.png)
35
 
36
- ![Example_3_C](https://cdn-uploads.huggingface.co/production/uploads/658814fd586088fd274d8cc1/njE7tHGxaJps7KyDmnKqY.png)
37
 
38
  ## Guide
39
  The full usage, installation guide and detailed workflow live in the [GitHub repo](https://github.com/SnJake/Ref2Font):
@@ -42,7 +46,7 @@ https://github.com/SnJake/Ref2Font
42
  ```
43
 
44
  ## What’s included
45
- - `Ref2FontV2.safetensors` (LoRA weights)
46
  - `Example Workflow/` (ComfyUI workflow with notes inside nodes)
47
  - `Example/` (input images and their atlases)
48
  - Post-processing scripts (`flux_pipeline.py`, `flux_grid_to_ttf.py`, `flux_upscale.py`)
@@ -77,9 +81,13 @@ https://huggingface.co/SnJake/Ref2Font/blob/main/Ref2FontV2.safetensors
77
  ```
78
  Place in: `ComfyUI/models/loras`
79
 
80
- ### ⚠️ IMPORTANT: Required Prompt
81
- To get the correct grid layout and character sequence, you **must** use this exact prompt:
82
- > Generate letters and symbols "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789!?.,;:-" in the style of the letters given to you as a reference.
 
 
 
 
83
 
84
  ### Input image rules
85
  - **Strict black & white only** (no gray, no shadows, no volume)
 
15
  - safetensors
16
  ---
17
 
18
+ # Ref2Font V3 — Contextual LoRA for FLUX.2 Klein 9B
19
 
20
+ **Ref2Font V3** is an updated **contextual LoRA** trained for **black-forest-labs/FLUX.2-klein-9B**. It generates **1280×1280 font atlases** from a single reference image "Aa".
21
+
22
+ ### Changes in V3:
23
+ - **Cyrillic Support:** Now generates Russian fonts using "Аа" as reference.
24
+ - **Extended Glyphs:** Added `"` and `&`.
25
+ - **Straighter Letters:** Improved alignment and reduced "jitter" in atlas generation.
26
 
27
  **Changes in V2:**
28
  - Fixed alignment issues (letters no longer "dance" or jump).
 
33
 
34
  ## Examples
35
 
36
+ ![Example_1_C](https://cdn-uploads.huggingface.co/production/uploads/658814fd586088fd274d8cc1/u7Jl9Un35cK1kLJty6I2g.png)
37
 
38
+ ![Example_3_C](https://cdn-uploads.huggingface.co/production/uploads/658814fd586088fd274d8cc1/IwlYEc154OqfAFdZaUz19.png)
 
 
39
 
40
+ ![Example_6_C](https://cdn-uploads.huggingface.co/production/uploads/658814fd586088fd274d8cc1/7Ky9ERLVmi4mRO-J9s5YV.png)
41
 
42
  ## Guide
43
  The full usage, installation guide and detailed workflow live in the [GitHub repo](https://github.com/SnJake/Ref2Font):
 
46
  ```
47
 
48
  ## What’s included
49
+ - `Ref2FontV3.safetensors` (LoRA weights)
50
  - `Example Workflow/` (ComfyUI workflow with notes inside nodes)
51
  - `Example/` (input images and their atlases)
52
  - Post-processing scripts (`flux_pipeline.py`, `flux_grid_to_ttf.py`, `flux_upscale.py`)
 
81
  ```
82
  Place in: `ComfyUI/models/loras`
83
 
84
+ ## ⚠️ IMPORTANT: V3 Required Prompts
85
+
86
+ **For Latin ("Aa" reference):**
87
+ > A technical font atlas grid of the Latin charset: "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789!?.,;:-"&". The style is strictly derived from the reference image "Aa".
88
+
89
+ **For Cyrillic ("Аа" reference):**
90
+ > A technical font atlas grid of the Cyrillic charset: "АБВГДЕЁЖЗИЙКЛМНОПРСТУФХЦЧШЩЪЫЬЭЮЯабвгдеёжзийклмнопрстуфхцчшщъыьэюя0123456789!?.,;:-"&". The style is strictly derived from the reference image "Аа".
91
 
92
  ### Input image rules
93
  - **Strict black & white only** (no gray, no shadows, no volume)