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 them in the Files & versions tab.