poluramus commited on
Commit
a2ebec0
·
verified ·
1 Parent(s): 4d86eed

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +15 -1
README.md CHANGED
@@ -6,10 +6,24 @@ This LoRA is designed to generate anime-style images.
6
  - **Training Data**: The model was trained on a small dataset of **13 images**.
7
  - **Training Steps**: **2250 steps** were used for training.
8
  - **Learning Rate**: Set to **2e-4**.
9
- - **Model**: Based on `black-forest-labs/FLUX.1-dev`, optimized for 8-bit mixed precision.
10
  - **LoRA Configuration**:
11
  - **Rank**: 32
12
  - **Batch Size**: 1
13
  - **Noise Scheduler**: `flowmatch`
14
  - **Resolution**: Images resized to 512, 768, and 1024.
15
  <br>
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6
  - **Training Data**: The model was trained on a small dataset of **13 images**.
7
  - **Training Steps**: **2250 steps** were used for training.
8
  - **Learning Rate**: Set to **2e-4**.
9
+ - **Model**: Based on `black-forest-labs/FLUX.1-dev`
10
  - **LoRA Configuration**:
11
  - **Rank**: 32
12
  - **Batch Size**: 1
13
  - **Noise Scheduler**: `flowmatch`
14
  - **Resolution**: Images resized to 512, 768, and 1024.
15
  <br>
16
+
17
+ # Anirand_v2_000002250
18
+ This LoRA is designed to generate anime-style images.
19
+ - **Training Framework**: Utilized `ai-toolkit` for the training process.
20
+ - **Training Data**: Same as Anirand_v1_000002250
21
+ - **Training Steps**: **2250 steps** were used for training.
22
+ - **Learning Rate**: Set to **2e-4**.
23
+ - **Model**: Based on `black-forest-labs/FLUX.1-dev`
24
+ - **LoRA Configuration**:
25
+ - **Rank**: 64
26
+ - **Batch Size**: 1
27
+ - **Noise Scheduler**: `flowmatch`
28
+ - **Resolution**: Images resized to 1024.
29
+ <br>