---
tags:
- text-to-image
- lora
- diffusers
- template:diffusion-lora
widget:
- text: '3D Sketchfab, A low-angle view of a brown military tank on a bright yellow background. The tank has a red number "88" on the side of it. There is a black X on the top of the tank. There are two black wheels on the bottom of it and a black stripe on the right side of the front of the left side of this tank.'
  output:
    url: images/S1.png
- text: '3D Sketchfab, An eye-level view of a small red building with a green sign on the front of it. The building is surrounded by a small patch of brown dirt. There is a brown roof on the right side of the building and a red door on the left side. There are small green plants on the ground around the building. A black telephone pole is in the center of the image. A white dish is on the roof of the house.'
  output:
    url: images/S2.png
- text: '3D Sketchfab, a vibrant orange box is adorned with a fish sculpture. The fish sculpture is a vibrant shade of blue, with a striped pattern on its body. It is positioned on a light purple background, with the fish head facing towards the right side of the frame. A knife with a black handle is positioned to the left of the fish sculpture, positioned on top of the box.'
  output:
    url: images/S3.png
base_model: black-forest-labs/FLUX.1-dev
instance_prompt: 3D Sketchfab
license: creativeml-openrail-m
---
# Castor-3D-Sketchfab-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-3D-Sketchfab-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           | 23 & 1.8k |
| Epoch                     | 10     | Save Every N Epochs       | 1      |

    Labeling: florence2-en(natural language & English)
    
    Total Images Used for Training : 39

## 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-3D-Sketchfab-Flux-LoRA"
trigger_word = "3D Sketchfab"  # 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     


## Trigger words 🧨

You should use `3D Sketchfab` to trigger the image generation.


## Download model

Weights for this model are available in Safetensors format.

[Download](/prithivMLmods/Castor-3D-Sketchfab-Flux-LoRA/tree/main) them in the Files & versions tab.