Fix Gradio version specification

README.md

@@ -1,78 +1,45 @@
 ---
 title: HarukiMix AI Image Generator
 emoji: 🎨
-colorFrom: purple
-colorTo: pink
+colorFrom: pink
+colorTo: purple
 sdk: gradio
-sdk_version: "4.26.0"
-python_version: "3.11"
+sdk_version: 4.20.0
 app_file: app.py
-pinned: false
+python_version: 3.11
 ---
 
-# 🎨 HarukiMix AI Image Generator
+# HarukiMix AI Image Generator
 
-Generate beautiful Japanese girl images using the harukimix model powered by Stable Diffusion XL.
+Beautiful Japanese girl image generator using the HarukiMix model.
 
 ## Features
 
-- **High-Quality Image Generation**: Uses the harukimix model, optimized for generating beautiful Japanese girl images
-- **Customizable Parameters**:
-  - Prompt and negative prompt control
-  - Adjustable inference steps (10-50)
-  - Guidance scale control (1.0-20.0)
-  - Custom image dimensions (512-1024px)
-  - Seed control for reproducibility
-- **Example Prompts**: Pre-loaded with 4 example prompts to get you started
-- **Fast Generation**: Optimized for quick image generation
-
-## Model Information
-
-- **Model**: harukimix (John6666/haruki-mix-v21-sdxl)
-- **Base**: Stable Diffusion XL
-- **Specialization**: Japanese girl character generation
-- **License**: OpenRAIL
+- Generate high-quality images of Japanese girls
+- Customizable prompts and negative prompts
+- Adjustable generation parameters
+- Multiple sample prompts for inspiration
 
 ## Usage
 
-1. Enter a text prompt describing the image you want to generate
-2. (Optional) Add a negative prompt to specify what to avoid
-3. Adjust parameters as needed:
-   - **Inference Steps**: Higher = better quality but slower (default: 30)
-   - **Guidance Scale**: How strictly to follow the prompt (default: 7.5)
-   - **Height/Width**: Image dimensions (default: 768x768)
-   - **Seed**: For reproducible results (-1 for random)
-4. Click "🎨 Generate Image"
-5. Wait for generation to complete
-
-## Generation Time
-
-- **First run**: 30-60 seconds (model loading)
-- **Subsequent runs**: 15-30 seconds (CPU) to 5-10 seconds (GPU)
-
-## Example Prompts
-
-Try these prompts to get started:
-
-- "beautiful japanese girl, long black hair, school uniform, smile, detailed face, masterpiece, best quality, 8k"
-- "cute japanese girl, pink hair, kawaii style, big eyes, happy expression, detailed, high quality"
-- "japanese woman, elegant kimono, traditional style, beautiful face, detailed, masterpiece"
-- "young japanese girl, casual clothes, natural lighting, smile, realistic, detailed, high quality"
+1. Enter your prompt describing the image you want to generate
+2. Optionally enter a negative prompt to exclude certain features
+3. Adjust the generation parameters (steps, guidance scale, etc.)
+4. Click "Generate Image" to create your image
 
-## Technical Details
+## Model
 
-- Built with Gradio for easy web interface
-- Uses PyTorch and Diffusers library
-- Supports both CPU and GPU inference
-- Automatic device detection (CUDA/CPU)
+This application uses the HarukiMix model, which is specifically trained to generate beautiful images of Japanese girls with high quality and detail.
 
-## Notes
+## Parameters
 
-- First generation may take longer as the model loads
-- The model is optimized for generating Japanese-style characters
-- Results vary based on prompt quality and parameters
-- Seed -1 generates random images; specific seeds produce reproducible results
+- **Prompt**: Description of the image to generate
+- **Negative Prompt**: Features to exclude from the image
+- **Steps**: Number of inference steps (higher = better quality but slower)
+- **Guidance Scale**: How much the model follows the prompt
+- **Height/Width**: Image dimensions
+- **Seed**: Random seed for reproducibility
 
 ## License
 
-This Space uses the OpenRAIL license family. See the model card for more details.
+This project uses the HarukiMix model which is licensed under the OpenRAIL license.