Qw-Sketch Open-source Image Generation Model - Free Generation of Sketch and Abstract-style Portrait Paintings

Qw Sketch

Developed by strangerzonehf

Qw-Sketch is a LoRA fine-tuned text-to-image diffusion model specializing in sketch and abstract-style portrait generation.

Image Generation Open Source License:Other #Sketch-style portrait generation #Vibrant background art creation #LoRA fine-tuned diffusion model

Downloads 22

Release Time : 1/20/2025

Model Overview

This model is fine-tuned with LoRA based on the black-forest-labs/FLUX.1-dev foundation model, capable of generating portrait images with unique artistic styles from text descriptions, particularly excelling in sketch and abstract expression styles.

Model Features

Artistic style generation

Capable of generating sketch and abstract-style portraits with unique artistic expression.

LoRA fine-tuning

LoRA fine-tuned on the FLUX.1-dev foundation model, enhancing specific style generation while preserving base model capabilities.

Detailed character description support

Can generate corresponding images based on detailed character descriptions (e.g., hairstyle, expression, clothing).

Model Capabilities

Text-to-image generation

Artistic style image generation

Portrait generation

Use Cases

Art creation

Portrait sketch creation

Generates sketch-style portraits from text descriptions

Examples showcase various styles of portrait sketches

Abstract art creation

Generates portraits with abstract expressionist styles

Examples include colorful abstract portraits

Concept design

Character concept design

Provides visual references for game or animation character designs

🚀 Qw-Sketch Model

This model focuses on text - to - image generation, leveraging LoRA and Diffusers technologies. It can generate vivid and diverse images based on given text descriptions, offering a unique image - generation experience.

🚀 Quick Start

Environment Setup

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/Qw-Sketch"
trigger_word = "Qw-Sketch"  
pipe.load_lora_weights(lora_repo)

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

Trigger Words

You should use Qw-Sketch to trigger the image generation.

Download Model

Weights for this model are available in Safetensors format. Download them in the Files & versions tab.

✨ Features

Text - to - Image Generation: Generate high - quality images according to text descriptions.
Rich Image Styles: Capable of generating various styles of images, including sketches and abstract paintings.

📦 Installation

The installation process mainly involves setting up the Python environment and loading the model. The code for setting up the environment is as follows:

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/Qw-Sketch"
trigger_word = "Qw-Sketch"  
pipe.load_lora_weights(lora_repo)

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

💻 Usage Examples

Basic Usage

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/Qw-Sketch"
trigger_word = "Qw-Sketch"  
pipe.load_lora_weights(lora_repo)

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

# Generate an image
text_prompt = 'Qw-Sketch, An eye-level sketch of a young Asian man with dark brown hair...'
image = pipe(text_prompt).images[0]
image.save('generated_image.png')

📚 Documentation

Model Description

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	16 & 2230
Epoch	17	Save Every N Epochs	1

Labeling: florence2 - en(natural language & English) Total Images Used for Training : 18

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

📄 License

This model is licensed under the flux - 1 - dev - non - commercial - license. You can view the full license details here.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご