TinyLLaVA / TinyLLaVA_Factory

A Framework of Small-scale Large Multimodal Models
https://arxiv.org/abs/2402.14289
Apache License 2.0
658 stars 68 forks source link

what the difference between v1 and v1.1? #19

Open Eric-is-good opened 8 months ago

Eric-is-good commented 8 months ago

image Your v1.1 model has shown some improvement compared to v1. May I ask if you have used more data to train or some special methods?

baichuanzhou commented 8 months ago

Sure, they correspond to TinyLLaVA-share-C-TL and TinyLLaVA-share-C-SLM in the paper. You can find them in the appendix.