|
--- |
|
tags: |
|
- text-to-image |
|
- lora |
|
- diffusers |
|
- template:diffusion-lora |
|
widget: |
|
- text: 'Inkk, A black and white portrait of a womans face. The womans head is facing the left side of the frame. Her hair is cut in a bun. Her eyes are wide open. Her eyebrows are black and her lips are painted black. Her mouth is painted white. Her nose is black. She has a black microphone in her mouth. The background is white.' |
|
output: |
|
url: images/1.png |
|
- text: 'Inkk, A black and white drawing of a mans face. The man has a black mustache that is trimmed in black. His eyes are blue and he has black hair. He is wearing a black collar with black stripes on it. He also has earphones in his ears. The background is white.' |
|
output: |
|
url: images/2.png |
|
- text: 'Inkk, A black and white monochromatic portrait of a womans face. The womans head is facing the left side of the frame, her hair cascades over her shoulders. She is wearing a black dress with a white stripe down the center of her neck. Her ear is encased in a silver earring. Her hair is pulled back in a ponytail, adding a pop of color to the scene. The background is a stark white, creating a stark contrast to the womans silhouette.' |
|
output: |
|
url: images/3.png |
|
base_model: black-forest-labs/FLUX.1-dev |
|
instance_prompt: Inkk |
|
license: apache-2.0 |
|
--- |
|
 |
|
|
|
<Gallery /> |
|
|
|
|
|
# Model description for Inkk |
|
|
|
Image Processing Parameters |
|
|
|
| Parameter | Value | Parameter | Value | |
|
|---------------------------|--------|---------------------------|--------| |
|
| LR Scheduler | constant | Noise Offset | 0.03 | |
|
| Optimizer | AdamW | Multires Noise Discount | 0.1 | |
|
| Network Dim | 64 | Multires Noise Iterations | 10 | |
|
| Network Alpha | 32 | Repeat & Steps | 19 & 2770 | |
|
| Epoch | 23 | Save Every N Epochs | 1 | |
|
|
|
Labeling: florence2-en(natural language & English) |
|
|
|
Total Images Used for Training : 55 |
|
|
|
## Best Dimensions & Inference |
|
|
|
| **Dimensions** | **Aspect Ratio** | **Recommendation** | |
|
|-----------------|------------------|---------------------------| |
|
| 1280 x 832 | 3:2 | Best | |
|
| 1024 x 1024 | 1:1 | Default | |
|
|
|
### Inference Range |
|
|
|
- **Recommended Inference Steps:** 30–35 |
|
|
|
## Setting Up |
|
```python |
|
import torch |
|
from pipelines import DiffusionPipeline |
|
|
|
base_model = "black-forest-labs/FLUX.1-dev" |
|
pipe = DiffusionPipeline.from_pretrained(base_model, torch_dtype=torch.bfloat16) |
|
|
|
lora_repo = "strangerzonehf/Inkk" |
|
trigger_word = "Inkk" |
|
pipe.load_lora_weights(lora_repo) |
|
|
|
device = torch.device("cuda") |
|
pipe.to(device) |
|
``` |
|
## Trigger words |
|
|
|
You should use `Inkk` to trigger the image generation. |
|
|
|
## Download model |
|
|
|
Weights for this model are available in Safetensors format. |
|
|
|
[Download](/strangerzonehf/Inkk/tree/main) them in the Files & versions tab. |
|
|
