---
tags:
- text-to-image
- lora
- diffusers
- template:diffusion-lora
widget:
- text: 'Qx-Art, A black and white drawing of a fox on a white paper. The fox has a long tail, a long snout, and a small black eye. There is a black string tied around the foxs neck, and the fox has two small black circles on the front of its head. There are two long legs on the left side of the fox, and one on the right side.'
  output:
    url: images/1.png
- text: 'Qx-Art, Captured from a high-angle perspective at eye-level on a stark white wall. The light bulb is composed of a thin black wire looped around the bulb. The bulb is adorned with a red fish-shaped object, adorned with two small white circles. The top of the bulb is encircled by a black wire, adding a pop of color to the otherwise monochromatic image.'
  output:
    url: images/2.png
- text: 'Qx-Art, A close-up view of a black wire pig head on a light pink wall. The pigs face is made up of two black circles, one on each side of the pigs mouth, and two on the other side of its nose. There are two small black circles on the top of its head. There is a black line running from the pig to the right of its mouth.'
  output:
    url: images/3.png
- text: 'Qx-Art, Captured from a high-angle perspective at eye-level on a beige canvas wall. The fish, painted in a vibrant shade of brown, is suspended from a black wire, suspended in the air by a metal hook. The hook is attached to the top of the net, adding a touch of depth to the image. The wall is adorned with a textured pattern of lines, creating a stark contrast to the fish.'
  output:
    url: images/4.png
- text: 'Qx-Art, A high-angle view of a light bulb made entirely of twisted black wire, hanging against a smooth gray wall. Inside the bulb is a small yellow star-shaped object, giving it a bright focal point. The wire extends from the bulbs base, looping upward in a spiral shape, creating a dynamic visual effect. Shadows from the wire pattern fall delicately onto the wall behind it.'
  output:
    url: images/5.png
- text: 'Qx-Art, A close-up of a wire elephant head mounted on a pale blue wall. The elephants trunk curls downward and ends in a spiral. Two black circles serve as its eyes, and its ears are outlined with thin, looped wire. A single black wire extends from the elephants trunk to the bottom right corner of the frame, creating a sense of movement.'
  output:
    url: images/6.png
base_model: black-forest-labs/FLUX.1-dev
instance_prompt: Qx-Art
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
---
![dzdfvbdfdfdbbdfvbszdfvbsad.png](https://cdn-uploads.huggingface.co/production/uploads/65bb837dbfb878f46c77de4c/se1gnwN60eWuCg4k2DWqE.png)

<Gallery />

# Model description for Qx-Art

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           | 18 & 2630 |
| Epoch                     |  17   | Save Every N Epochs       | 1    |

    Labeling: florence2-en(natural language & English)
    
    Total Images Used for Training : 19

## 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/Qx-Art"
trigger_word = "Qx-Art"  
pipe.load_lora_weights(lora_repo)

device = torch.device("cuda")
pipe.to(device)
```
## Trigger words

You should use `Qx-Art` to trigger the image generation.

## Download model

Weights for this model are available in Safetensors format.

[Download](/strangerzonehf/Qx-Art/tree/main) them in the Files & versions tab.