Castor-Collage-Dim-Flux-LoRA

Prompt

A vibrant red spiderman is depicted in a collage of multiple panels. The background is a vibrant orange, adding a pop of color to the scene. The spidermans suit is a deep red, with a white stripe running down the center of his chest. His eyes are large, white, and his mouth is slightly open. His hands are positioned in a way that he is facing towards the right side of the frame. The entire frame is black, creating a striking contrast with the red and orange background.

Prompt

A vibrant, black-suited Batman is depicted in a collage of multiple panels. The background is a deep, intense blue, adding a dramatic tone to the scene. Batmans suit is a rich black with a dark gray bat symbol on his chest. His eyes are narrow and white, and his mouth is set in a serious expression. His cape flows around him, and he is facing toward the right side of the frame, giving a sense of movement. The entire frame is outlined in dark gray, creating a bold contrast with the dark suit and blue background.

Prompt

A vibrant, red-suited Deadpool is depicted in a collage of multiple panels. The background is a bold, electric yellow, adding an energetic vibe to the scene. Deadpool’s suit is a striking red with black patches around his eyes and across his torso, creating a dynamic look. His eyes are expressive and white, and his mouth is curved into a mischievous smirk. His hands are positioned confidently, holding one of his katanas, and he is facing toward the right side of the frame, exuding a playful, yet intense energy. The entire frame is outlined in black, creating a sharp contrast with the red suit and vibrant yellow background.

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-Collage-Dim-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	28 & 3.2k
Epoch	12	Save Every N Epochs	1

Labeling: florence2-en(natural language & English)

Total Images Used for Training : 27

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-Collage-Dim-Flux-LoRA"
trigger_word = "collage"  # Leave trigger_word blank if not used.
pipe.load_lora_weights(lora_repo)

device = torch.device("cuda")
pipe.to(device)

Sample Image + Prompt

Prompt

A dynamic scene from GTA (Grand Theft Auto) is depicted in a collage of multiple panels. The background is a gritty, urban cityscape with neon lights casting vibrant shades of purple, blue, and orange, capturing the chaotic energy of the city at night. A character in casual streetwear, holding a bag of cash in one hand and a weapon in the other, stands confidently, slightly facing the right side of the frame. The character has a focused expression, exuding a sense of rebellion and risk. The entire frame is outlined in dark gray, creating a bold contrast with the vibrant city lights and shadowed streets, evoking the intense, action-packed vibe of GTA.

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 🧨

Trigger words: You should use collage to trigger the image generation.

Download model

Weights for this model are available in Safetensors format.

Download them in the Files & versions tab.