Hugging Face's logo

Spaces:
yeq6x
/
QIE-Image2GuideBody
Running on Zero

App Files Community
QIE-Image2GuideBody / README.md
yeq6x's picture
yeq6x
Add default weight configuration for Stage 1 and Stage 2 LoRAs in app.py, and update README.md to document new environment variables for customization.
826fa01
preview code
|
raw
history blame contribute delete
2.31 kB

A newer version of the Gradio SDK is available: 6.3.0

Upgrade
metadata
title: QIE-Image2GuideBody
emoji: 🎨✨
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 6.2.0
app_file: app.py
pinned: false
license: apache-2.0
short_description: Anime character to guide body converter

🎨✨ QIE-Image2GuideBody

A single-pass image converter that applies multiple LoRAs simultaneously to transform anime character images.

Overview

This application performs conversion in a single inference pass by combining multiple LoRA models:

Combined LoRAs:

  • Lightning LoRA: Enables fast 4-step inference
  • Stage1 LoRA: Custom transformation effect 1
  • Stage2 LoRA: Custom transformation effect 2

All three LoRAs are applied simultaneously with equal weights (1.0, 1.0, 1.0) to produce the final result.

Technology

Built on Qwen-Image-Edit-2511 with:

  • Multiple LoRA combination support
  • Qwen-Image-Lightning-2511 for fast 4-step inference
  • Custom LoRA models applied simultaneously

Configuration

The application uses environment variables for customization:

Model Settings

  • BASE_MODEL: Base model repository (default: Qwen/Qwen-Image-Edit-2511)

LoRA Settings

  • LIGHTNING_LORA_REPO: Lightning 4-step LoRA repository (default: lightx2v/Qwen-Image-Edit-2511-Lightning)
  • LIGHTNING_LORA_WEIGHT: Lightning LoRA weight filename (default: Qwen-Image-Edit-2511-Lightning-4steps-V1.0-bf16.safetensors)
  • STAGE1_LORA_REPO: Repository for Stage 1 LoRA
  • STAGE1_LORA_WEIGHT: Weight filename for Stage 1 LoRA
  • STAGE1_WEIGHT_DEFAULT: Default weight value for Stage 1 LoRA (default: 1.0)
  • STAGE2_LORA_REPO: Repository for Stage 2 LoRA
  • STAGE2_LORA_WEIGHT: Weight filename for Stage 2 LoRA
  • STAGE2_WEIGHT_DEFAULT: Default weight value for Stage 2 LoRA (default: 1.0)

Prompt Settings

  • STAGE1_PROMPT: Prompt used for inference (default: "Convert anime character to base body structure")
  • STAGE2_PROMPT: Not used in current implementation (reserved for future use)

Usage

  1. Upload an anime character image
  2. Click "Generate"
  3. View the result with all LoRAs applied simultaneously

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference