Upload 3 files

README.md

@@ -1,103 +1,78 @@
----
-title: Stable Diffusion Style Explorer
-emoji: 🎨
-colorFrom: purple
-colorTo: pink
-sdk: gradio
-sdk_version: 5.9.0
-app_file: app.py
-pinned: false
-license: mit
----
-
-# 🎨 Stable Diffusion Guided Style Explorer
-
-This project demonstrates the power of **Textual Inversion** combined with **Artistic Guided Sampling** to steer Stable Diffusion v1-4 towards specific artistic styles and visual attributes.
-
-## 🚀 Overview
-
-The project consists of two main components:
-1.  **Jupyter Notebook (`stable_conceptualizer_inference (1).ipynb`)**: A research-focused environment for experimenting with loss-based guidance and exploring the mechanics of Textual Inversion.
-2.  **Gradio Web App (`app.py`)**: A user-friendly interface for generating images with real-time control over artistic guidance.
-
----
-
-## 🔬 Core Concepts
-
-### 1. Textual Inversion
-Instead of retraining the entire Stable Diffusion model, we use **Textual Inversion** to learn new "pseudo-words" (tokens) that represent specific concepts or styles.
-*   **Token Injection**: By adding a learned embedding to the CLIP text encoder, we can invoke complex styles with a simple keyword like `<style>`.
-*   **Integrated Styles**: The project includes styles from the SD Concepts Library such as:
-    *   **Madhubani Art**: Traditional Indian folk art.
-    *   **Cat Toy**: Plastic, cute aesthetic.
-    *   **Seletti**: Porcelain/ceramic design.
-    *   **Indian Watercolor**: Expressive portraits.
-    *   **Chucky & Anime Boy**: Character-specific styles.
-
-### 2. Guided Sampling (Artistic Steering)
-While standard sampling follows the prompt, our **Guided Sampling** implementation injects extra gradients during the diffusion process to maximize specific visual features:
-
-| Guidance Type | Technical Implementation | Artistic Effect |
-| :--- | :--- | :--- |
-| **Contrast** | Maximizes pixel variance from 0.5 | Dramatic lighting, deep shadows. |
-| **Complexity** | Edge-detection (Sobel-like) gradients | Intricate details, sharp textures. |
-| **Vibrancy** | Maximizes color channel variance | Vivid, punchy color saturation. |
-
----
-
-## 🛠 Features
-
-### 📓 Jupyter Notebook
-*   **Step-by-Step implementation**: Detailed code for applying loss gradients during the UNet sampling loop.
-*   **Reproducibility**: Uses fixed seeds (`torch.Generator`) to isolate the effect of guidance scales, ensuring that changes in output are purely due to artistic steering.
-*   **Deep Dive Documentation**: Explains the math and logic behind each loss function.
-
-## 📱 User Guide: Gradio App
-
-The **Stable Diffusion Style Explorer** provides two distinct ways to interact with the model.
-
-### 1. Single Style Generation
-Use this tab for granular control over a specific aesthetic.
-*   **Select Style**: Choose one of the 5+ pre-loaded textual inversion concepts.
-*   **Prompt**: Write your prompt using the `<style>` placeholder. (e.g., `"a futuristic city in the style of <style>"`).
-*   **Artistic Sliders**:
-    *   **Contrast Strength**: Increase to add drama and deeper shadows.
-    *   **Complexity Strength**: Increase to force intricate patterns and fine details.
-    *   **Vibrancy Strength**: Increase for more saturated and "glowing" colors.
-*   **Seed Management**: Each style comes with a pre-configured "best" seed, but you can override it to explore different variations.
-
-### 2. Compare All Styles
-Use this tab to see how a single prompt manifests across different artistic interpretations.
-*   **Batch Generation**: Generates 3 images (default) for every single style simultaneously.
-*   **Unified Guidance**: Apply the same Contrast, Complexity, and Vibrancy scales across all styles to compare their response to guidance.
-*   **Style-Specific Seeds**: Configure individual seeds for each style to ensure reproducibility across separate runs.
-
----
-
-## 🚦 Getting Started
-
-### Prerequisites
-*   Python 3.10+
-*   GPU with 8GB+ VRAM (Recommended)
-*   Hugging Face Token (for model and style loading)
-
-### Setup
-1.  **Clone the repository**.
-2.  **Environment Setup**:
-    ```bash
-    pip install -r requirements.txt
-    ```
-3.  **Run the App**:
-    ```bash
-    python app.py
-    ```
-
-## 💡 Tips for Best Results
-*   **Guidance Scales**: Typical effective values for Contrast/Complexity/Vibrancy range from **500 to 1500**. Start low and increase gradually.
-*   **Prompting**: Keep prompts relatively simple to let the Textual Inversion style shine.
-*   **Seeds**: If you find an image layout you like, keep the seed fixed while adjusting the loss sliders to see exactly how the guidance "sculpts" that specific composition.
-
-## 📜 Credits
-*   **Model**: Stable Diffusion v1-4
-*   **Concepts**: 🤗 Hugging Face [SD Concepts Library](https://huggingface.co/sd-concepts-library)
-*   **Implementation**: Custom Triple-Loss Guidance Suite.
+---
+title: Stable Diffusion Style Explorer
+emoji: 🎨
+colorFrom: purple
+colorTo: pink
+sdk: gradio
+sdk_version: 6.0.0
+app_file: app.py
+pinned: false
+license: mit
+---
+
+# Stable Diffusion Style Explorer
+
+An interactive web application for exploring different artistic styles using Stable Diffusion with textual inversion.
+
+## Features
+
+- **5 Pre-configured Styles**: Cat Toy, GTA5 Artwork, Birb Style, Midjourney Style, and Arcane Style
+- **Single Style Mode**: Generate images with a specific style, custom seed, and parameters
+- **Compare All Styles**: Generate the same prompt across all 5 styles simultaneously
+- **Seed Control**: Full control over random seeds for reproducible results
+- **Adjustable Parameters**: Configure inference steps and guidance scale
+
+## Usage
+
+### Single Style Mode
+1. Enter your prompt (use `<style>` as a placeholder for the style token)
+2. Select a style from the dropdown
+3. Set your desired seed value
+4. Adjust inference steps and guidance scale if needed
+5. Click "Generate Image"
+
+### Compare All Styles Mode
+1. Enter your prompt (use `<style>` as a placeholder)
+2. Set a base seed value
+3. Each style will use: `base_seed + (style_index * 100)`
+4. Click "Generate All Styles" to see all variations
+
+## Styles
+
+- **Cat Toy**: Cute cat toy aesthetic
+- **GTA5 Artwork**: GTA V game art style
+- **Birb Style**: Artistic bird illustration style
+- **Midjourney Style**: Midjourney AI art aesthetic
+- **Arcane Style**: Arcane Netflix series art style
+
+## Technical Details
+
+- **Base Model**: CompVis/stable-diffusion-v1-4
+- **Textual Inversion**: Concepts from [SD Concepts Library](https://huggingface.co/sd-concepts-library)
+- **Framework**: Gradio + Diffusers
+- **GPU**: Recommended for faster generation
+
+## Local Development
+
+```bash
+# Clone the repository
+git clone <your-repo-url>
+cd <repo-name>
+
+# Install dependencies
+pip install -r requirements.txt
+
+# Run the app
+python app.py
+```
+
+## Deployment to Hugging Face Spaces
+
+1. Create a new Space on Hugging Face
+2. Select "Gradio" as the SDK
+3. Upload `app.py`, `requirements.txt`, and `README.md`
+4. The app will automatically build and deploy
+
+## License
+
+MIT License - Feel free to use and modify!

app.py

@@ -544,6 +544,5 @@ with gr.Blocks(title="Stable Diffusion Style Explorer",theme=gr.themes.Soft()) a
 if __name__ == "__main__":
     #demo.launch(server_name="0.0.0.0", server_port=7860, share=True)
     print("RUNNING THIS FILE:", __file__)
-    python -c "import gradio; print(gradio.__version__)"
     demo.launch()
 

requirements.txt

@@ -1,4 +1,4 @@
-gradio>=5.9.0
+gradio>=6.0.0
 torch
 torchvision
 diffusers