# Small-scale Multimodal

TinyLLaVA-Phi-2-SigLIP-3.1B is a small-scale large multimodal model with 3.1B parameters, combining the Phi-2 language model and SigLIP vision model, outperforming some 7B models.

TinyLLaVA is a small-scale large multimodal model framework that significantly reduces the number of parameters while maintaining high performance. The 3.1B version outperforms similar 7B-scale models in multiple benchmarks.