A simple Streamlit dashboard that uses Gemma 3:4b via Ollama to identify car brands and colors from uploaded images.
- 📸 Image Upload: Drag & drop car images (JPG, PNG)
- 🤖 AI Analysis: Powered by Gemma 3:4b model
- 🎨 Structured Output: Returns car brand and color only
- ⚡ Real-time Processing: Instant results via localhost
-
Install dependencies:
pip install streamlit openai pillow
-
Start Ollama server:
ollama serve
-
Pull the model:
ollama pull gemma3:4b
-
Run the dashboard:
streamlit run gemma_dashboard.py
gemma_vehicle.py- Car expert agent with image analysisgemma_dashboard.py- Streamlit web interfaceLICENSE- Apache 2.0 license
- Python 3.8+
- Ollama running on
localhost:11434 - Gemma 3:4b model installed
- Use clear, well-lit car images
- Avoid heavily modified vehicles
- Works best with standard car brands
© 2025 Rebels Software sp. z o.o. | rebels.software
Licensed under Apache 2.0