Merge branch 'pr/1' into main - resolve conflicts keeping updated versions

README.md

@@ -122,3 +122,115 @@ This project is for research and educational purposes. Medical applications shou
 ## Support
 
 For issues or questions, please refer to the Hugging Face Space documentation or create an issue in the project repository.
+=======
+---
+title: Medical Image Analysis Tool
+emoji: 🏥
+colorFrom: blue
+colorTo: green
+sdk: gradio
+sdk_version: 5.49.1
+app_file: app.py
+pinned: false
+license: mit
+---
+
+# 🏥 Medical Image Analysis Tool
+
+An AI-powered medical image analysis application using advanced detection models and large language models for medical image interpretation.
+
+## Features
+
+- **Advanced Object Detection**: Uses RF-DETR (Real-time Fine-grained Detection Transformer) for precise object detection
+- **Medical AI Analysis**: Integrates MedGemma, a specialized medical vision-language model
+- **Interactive Interface**: Built with Gradio for easy web-based interaction
+- **Configurable Thresholds**: Adjustable confidence thresholds for detection sensitivity
+- **Model Size Selection**: Choose between MedGemma 4B (faster) or 27B (more accurate) models
+- **GPU Acceleration**: Optimized for GPU usage when available with 4-bit quantization
+- **Automatic Model Downloads**: Models download automatically from Hugging Face Hub
+
+## Models Used
+
+- **RF-DETR Medium**: State-of-the-art object detection model
+- **MedGemma 4B/27B**: Medical-specialized vision-language models for analysis and descriptions
+  - 4B model: Faster inference, lower memory usage
+  - 27B model: Higher accuracy, requires more resources
+
+## Usage
+
+1. **Upload Image**: Click on the image upload area or drag and drop a medical image
+2. **Adjust Settings**:
+   - Use the confidence threshold slider to control detection sensitivity
+   - Select model size (4B for speed, 27B for accuracy)
+3. **Analyze**: Click "Analyze Image" to run the AI analysis
+4. **View Results**: See the annotated image with detected objects and AI-generated descriptions
+
+## Installation & Setup
+
+This application is designed to run on Hugging Face Spaces. The following files are required:
+
+- `app.py` - Main application file (optimized for Spaces)
+- `requirements.txt` - Python dependencies
+- `packages.txt` - System packages
+- `README.md` - This documentation
+
+## Model Loading
+
+**RF-DETR Model:**
+- Upload your trained `rf-detr-medium.pth` file to the Space
+- The application will automatically find and load it
+
+**MedGemma Models:**
+- Models download automatically from Hugging Face Hub on first use
+- No manual installation required
+- Choose between 4B (faster) or 27B (more accurate) models
+
+## Space Configuration
+
+For optimal performance, configure your Space settings:
+- **Hardware**: GPU (T4 minimum, A100 recommended for 27B models)
+- **Storage**: Enable persistent storage for model caching
+- **Timeout**: 30+ minutes for large model downloads
+
+## Technical Details
+
+- **Framework**: PyTorch + Transformers
+- **Interface**: Gradio
+- **Computer Vision**: OpenCV, PIL, Supervision
+- **Hardware**: Optimized for both CPU and GPU inference
+
+## Performance Tips
+
+- **Model Selection**: Use MedGemma 4B for faster processing or 27B for higher accuracy
+- **Confidence Thresholds**: Higher values reduce false positives but may miss subtle findings
+- **GPU Acceleration**: The application automatically uses GPU acceleration when available
+- **Memory Optimization**: Uses 4-bit quantization to reduce memory usage
+- **Model Caching**: Models are cached after first load for faster subsequent analyses
+
+## Limitations
+
+- Requires significant computational resources for optimal performance
+- Best suited for medical imaging applications
+- Results should be verified by qualified medical professionals
+
+## Development
+
+To run locally:
+
+```bash
+pip install -r requirements.txt
+python app.py
+```
+
+**Note**: For local development, you'll need to:
+1. Install the RF-DETR package or ensure it's available
+2. Place your `rf-detr-medium.pth` file in the project directory
+3. Models will download automatically on first run
+
+## License
+
+This project is for research and educational purposes. Medical applications should be developed and validated according to appropriate regulatory standards.
+
+## Support
+
+For issues or questions, please refer to the Hugging Face Space documentation or create an issue in the project repository.

app.py

@@ -90,25 +90,8 @@ memory_manager = MemoryManager()
 
 def find_checkpoint() -> Optional[str]:
     """Find RF-DETR checkpoint in various locations."""
-    # Check for HuggingFace model repository first (recommended)
-    import os
-    hf_model_id = os.environ.get("RFDETR_HF_REPO")
-    if hf_model_id:
-        try:
-            from huggingface_hub import hf_hub_download
-            print(f"Downloading RF-DETR from HuggingFace Hub: {hf_model_id}")
-            checkpoint_path = hf_hub_download(
-                repo_id=hf_model_id,
-                filename="rf-detr-medium.pth",
-                cache_dir="/tmp/hf_cache"
-            )
-            return checkpoint_path
-        except Exception as e:
-            print(f"Failed to download from HF Hub: {e}")
-
-    # Fallback to local files
     candidates = [
-        "rf-detr-medium.pth",  # Current directory (direct upload)
+        "rf-detr-medium.pth",  # Current directory
         "/tmp/results/checkpoint_best_total.pth",
         "/tmp/results/checkpoint_best_ema.pth",
         "/tmp/results/checkpoint_best_regular.pth",

packages.txt

@@ -2,7 +2,7 @@ libgl1-mesa-glx
 libglib2.0-0
 libsm6
 libxext6
-libxrender1
+libxrender
 libgomp1
 ffmpeg
 build-essential

requirements.txt

@@ -1,11 +1,11 @@
-torch>=2.0.0
-transformers>=4.30.0
-gradio>=4.0.0
-pillow>=10.0.0
-opencv-python>=4.8.0
-supervision>=0.18.0
-psutil>=5.9.0
-numpy>=1.24.0
-imageio>=2.31.0
-imageio-ffmpeg>=0.4.8
-requests>=2.31.0
+torch
+transformers>
+gradio
+pillow
+opencv-python
+supervision>
+psutil
+numpy
+imageio
+imageio-ffmpeg
+requests