1min snip

Latent Space: The AI Engineer Podcast — Practitioners talking LLMs, CodeGen, Agents, Multimodality, AI UX, GPU Infra and all things Software 3.0 cover image

How to train your own Large Multimodal Model — with Hugo Laurençon & Leo Tronchon of HuggingFace M4

Latent Space: The AI Engineer Podcast — Practitioners talking LLMs, CodeGen, Agents, Multimodality, AI UX, GPU Infra and all things Software 3.0

NOTE

Benefits of Pre-trained Vision Encoder Trained with Text Objective for Multi-modal Models

Using a pre-trained vision encoder trained with text objective, such as contrastive loss like the clip loss, is shown to be better for building multi-modal models compared to using vision encoders trained only on classification or mini-modal tasks. The embeddings from a vision encoder trained with text can be more effective when integrated into a vision language model. Researchers have demonstrated that integrating a vision encoder trained with text objective through lightweight updates, such as adding a linear layer trained on top of the encoder output, can be effective.

00:00

Get the Snipd
podcast app

Unlock the knowledge in podcasts with the podcast player of the future.
App store bannerPlay store banner

AI-powered
podcast player

Listen to all your favourite podcasts with AI-powered features

Discover
highlights

Listen to the best highlights from the podcasts you love and dive into the full episode

Save any
moment

Hear something you like? Tap your headphones to save it with AI-generated key takeaways

Share
& Export

Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more

AI-powered
podcast player

Listen to all your favourite podcasts with AI-powered features

Discover
highlights

Listen to the best highlights from the podcasts you love and dive into the full episode