deepvk/llava-saiga-8b

Image-Text-to-Text • Updated Aug 13 • 75 • 14

Note VLM model based on Saiga-8b and tuned in LLaVA setup

deepvk/llava-gemma-2b-lora

Image-Text-to-Text • Updated Aug 13 • 109 • 8

Note VLM model based on Gemma-2b and tuned with LoRA in LLaVA setup

Note GPT-based instruction dataset for LLaVA-style training

Note Translated version of GQA benchmark

Note Translated version of MMBench benchmark