multimodal

MM1 : Apple’s Multimodal LLM pretraining revealed

On 14th March 2024, a team of researchers from Apple released Multimodal LLM MM1, their findings on Multimodal LLM Pre-training and fine-tuning. In this work the importance of various architecture components and data choices in Large Language Models has been studied. Through thorough analysis of the image encoder, vision language connector and different pre-training data options, several […]

benchmark

Gemma : The secret ingredient behind gemini?

Gemma : On Feb 21, 2024, Google announced a new series of light weight state of the art open models named Gemma built from the same research and technology used to create gemini models. In this blog we will see a quick performance summary of the Gemma, followed by a quick review of its key components

Scroll to Top