Overview
Meta has released Llama 3.2, the first open-source Llama model with vision capabilities, enabling developers to build multimodal applications.
Model Variants
| Model | Parameters | Vision | Use Case |
|---|---|---|---|
| Llama 3.2 1B | 1B | No | Edge devices |
| Llama 3.2 3B | 3B | No | Mobile |
| Llama 3.2 11B | 11B | Yes | Multimodal |
| Llama 3.2 90B | 90B | Yes | Advanced tasks |
Key Features
- Vision Understanding: Process images and text
- Edge Ready: Small models for mobile
- Open Source: Fully open weights
- Commercial Use: Permissive license
Resources
Source: Meta