# SmolVLM2-2.2B-Instruct
A compact 2.2B parameter vision-language model from HuggingFace with video understanding capabilities.
## Key Features
- **Compact & Fast** - Only 2.2B parameters, runs efficiently on consumer hardware
- **Vision & Video** - Understands images and video frames
- **Instruction-tuned** - Optimized for following user instructions
- **Apache 2.0** - Fully open source
## Quick Start
```bash
ollama run richardyoung/smolvlm2-2.2b-instruct
```
## Available Formats
| Format | Size | Description |
|--------|------|-------------|
| Q4_K_M (default) | 1.0 GB | Best balance of quality and speed |
| Q8_0 | 1.8 GB | Higher quality |
| F16 | 3.4 GB | Full precision |
## Technical Requirements
- **Minimum:** 4GB RAM, any modern CPU
- **Recommended:** 8GB RAM or Apple Silicon Mac
## Links
- **Original Model:** HuggingFaceTB/SmolVLM2-2.2B-Instruct
- **GGUF Quantizations:** richardyoung/SmolVLM2-2.2B-Instruct-GGUF
## Credits
- **Original Model:** HuggingFace
- **Quantization:** Richard Young (deepneuro.ai/richard)
## License
Apache 2.0