---
tags:
- text-to-image
- lora
- diffusers
- template:diffusion-lora
widget:
- text: 'happy halloween, An animated image of a black bat sitting on top of a jack-o-lantern. The background is a vibrant orange, and there are white ghosts in the background. Above the bat, there is a text that reads "HAPPY HALLOWEEN" in black letters. The bat has a black face with yellow eyes and a black tail.'
  output:
    url: images/hw1.webp
- text: 'happy halloween, Captured at eye-level on a vibrant day, a spooky Halloween scene features a jack-o-lantern in the center of the frame, adorned with a pointed black hat. The pumpkins face is glowing with glowing orange lights, adding a touch of warmth to the scene. The scene is set in a field of tall, dry grass, with tall twigs sticking out of the ground. In the background, a forest of tall trees can be seen, adding depth to the composition.'
  output:
    url: images/hw2.webp
- text: 'At the edge of a foggy graveyard, a Halloween scene unfolds with a lone carved pumpkin resting on a stone bench, its face glowing with flickering candlelight. The pumpkin sits beside a cluster of dried flowers, and a ghostly white sheet flutters from a nearby tree branch. A row of aged tombstones stretches into the background, partially hidden by the mist that blankets the ground, giving the scene an eerie and timeless atmosphere.'
  output:
    url: images/hw3.webp
base_model: black-forest-labs/FLUX.1-dev
instance_prompt: happy halloween
license: creativeml-openrail-m
---
# Castor-Happy-Halloween-Flux-LoRA

<Gallery />

**The model is still in the training phase. This is not the final version and may contain artifacts and perform poorly in some cases.**

## Model description 

**prithivMLmods/Castor-Happy-Halloween-Flux-LoRA**

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           | 20 & 1400 |
| Epoch                     | 12     | Save Every N Epochs       | 1      |

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

## Setting Up
```
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 = "prithivMLmods/Castor-Happy-Halloween-Flux-LoRA"
trigger_word = "happy halloween"  # Leave trigger_word blank if not used.
pipe.load_lora_weights(lora_repo)

device = torch.device("cuda")
pipe.to(device)
```
## App File Structure

/project-root/

├── .gitattributes         
├── README.md      
├── app.py             
├── pythonproject.py     

# Best Dimensions

- 512 X 512
- 1024 X 1024
- 768 X 1024

## Trigger words 🧨

> [!WARNING]
> **Trigger words:**  You should use `happy halloween` to trigger the image generation.

## Download model

Weights for this model are available in Safetensors format.

[Download](/prithivMLmods/Castor-Happy-Halloween-Flux-LoRA/tree/main) them in the Files & versions tab.