Spaces:
Sleeping
Sleeping
File size: 2,256 Bytes
1221014 ffe870b 1221014 ffe870b 1221014 e2750f5 1221014 e2750f5 1221014 e2750f5 1221014 ffe870b e2750f5 4bd7b20 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 |
# CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
## Project Overview
This is a Hugging Face Spaces application that generates images in Pablo Picasso's cubist style using the FLUX.1-dev model from Black Forest Labs. The app uses Gradio to create a web interface with ZeroGPU acceleration for high-quality text-to-image generation.
## Architecture
- **app.py**: Gradio application using FLUX.1-dev for Picasso-style image generation
- **README.md**: Hugging Face Spaces configuration file with metadata (title, emoji, SDK version, etc.)
- **requirements.txt**: Python dependencies including PyTorch, Diffusers, and Gradio
- **.gitattributes**: Git LFS configuration for handling large model files and binary assets
## Key Commands
### Running the Application
```bash
python app.py
```
This launches the Gradio interface locally for testing (requires GPU).
### Hugging Face Spaces Deployment
The app is automatically deployed to Hugging Face Spaces with ZeroGPU. Changes to `app.py` will trigger redeployment.
## Development Notes
- Uses **FLUX.1-dev** model from Black Forest Labs for high-quality text-to-image generation
- **@spaces.GPU decorators** are required for ZeroGPU compatibility
- Model runs in **bfloat16** precision for optimal performance on H200 GPUs
- Output resolution is **1024x1024** pixels
- Optimal parameters: 28 inference steps, guidance scale 3.5
- Prompt enhancement automatically adds Picasso-style descriptors
## Code Architecture
### Core Components
- **Model Loading**: `load_model()` function initializes FLUX.1-dev pipeline with bfloat16 precision
- **Image Generation**: `generate_image()` function handles prompt enhancement and image generation with configurable parameters
- **UI Interface**: Gradio Blocks layout with prompt input, parameter sliders, and example buttons
- **Example System**: Pre-defined example prompts accessible through vertical button layout
### Key Functions
- `load_model()` (app.py:9): Loads and configures the FLUX.1-dev pipeline
- `generate_image()` (app.py:27): Main generation function with Picasso-style prompt enhancement
- Example setter functions (app.py:88-97): Populate prompt input with predefined examples |