MMMModal -- Multi-Images Multi-Audio Multi-turn Multi-Modal
This represents a significant advancement for the Malaysian context and beyond by enabling multimodal interactions in multiple languages.
The paper tackles the problem of multimodal comprehension by introducing a bilingual model that processes multi-images, multi-audio, and their combinations in multi-turn sessions, achieving proficiency in English and Malay with versions having 1.1B and 7B parameters.
Our contribution introduces a groundbreaking multimodal large language model designed to comprehend multi-images, multi-audio, and multi-images-multi-audio within a single multiturn session. Leveraging state-of-the-art models, we utilize the SigLIP encoder for visual inputs and the Whisper Encoder for audio inputs. Notably, this multimodal large language model is bilingual, proficient in understanding both English and Malay simultaneously. We proudly unveil two versions of this model: TinyLlama with 1.1B parameters, and Mistral with 7B parameters. With its ability to navigate diverse modalities and languages, our model represents a significant advancement for the Malaysian context and beyond. All models released at https://huggingface.co/collections/mesolitica/multimodal-malaysian-llm-65c6f893e03f78fa9e5c8859