 |
ML Newsnew | past | best | rss | submit | about |
| MM1.5: Methods, Analysis and Insights from Multimodal LLM Fine-Tuning (arxiv.org) | |
| 5 points by dailcooper on Oct 2, 2024 | hide | past | pdf | discuss on HN | |
|
In plain words: A family of vision-and-text models is trained with carefully mixed data at every stage, including text read from images and machine-written captions, to handle text-heavy pictures and several images. That data care lets the smallest 1B version perform strongly, rather than needing huge models. Abstract · MM1.5: Methods, Analysis & Insights from Multimodal LLM Fine-tuningWe present MM1.5, a new family of multimodal large language models (MLLMs) designed to enhance capabilities in text-rich image understanding, visual referring and grounding, and multi-image reasoning. Building upon the MM1 architecture, MM1.5 adopts a data-centric approach to model training, systematically exploring the impact of diverse data mixtures across the entire model training lifecycle. This includes high-quality OCR data and synthetic captions for continual pre-training, as well as an optimized visual instruction-tuning data mixture for supervised fine-tuning. Our models range from 1B to 30B parameters, encompassing both dense and mixture-of-experts (MoE) variants, and demonstrate that careful data curation and training strategies can yield strong performance even at small scales (1B and 3B). Additionally, we introduce two specialized variants: MM1.5-Video, designed for video understanding, and MM1.5-UI, tailored for mobile UI understanding. Through extensive empirical studies and ablations, we provide detailed insights into the training processes and decisions that inform our final designs, offering valuable guidance for future research in MLLM development. | |
| add comment on HN |