Quick start
ollama run llava-llama3Available sizes
| Tag | Size | Quantization | Context | Min RAM |
|---|---|---|---|---|
| llava-llama3:latest | 5.5GB | q4_k_m | 8K context | 6.9 GB |
Strengths & Limitations
Strengths
- Strong performance on benchmarks
- Built on Llama 3 Instruct
- Multimodal capabilities
Related models
llavaMultimodal
🌋 LLaVA is a novel end-to-end trained large multimodal model that combines a vision encoder and Vicuna for general-purpose visual and language understanding. Updated to version 1.6.
12.9M pullsminicpm-vMultimodal
A series of multimodal LLMs (MLLMs) designed for vision-language understanding.
4.6M pullsqwen3-vlMultimodal
The most powerful vision-language model in the Qwen model family to date.
1.6M pullsqwen2.5vlMultimodal
Flagship vision-language model of Qwen and also a significant leap from the previous Qwen2-VL.
1.3M pulls