Tag
Every Latent Daily story, tool, model and resource tagged “Multimodal”.
Meta published Muse Spark Open, a mixture-of-experts model with roughly 816B parameters, 42B active per token, multimodal input and a 1M-token context window.
AI Desk1 min read
CLIP learned from 400 million image-caption pairs to match pictures with text, enabling zero-shot image classification and powering text-to-image generators.
AI Desk4 min read
The 2023 LLaVA paper connected a vision encoder to an open language model and trained it on AI-generated visual instructions, creating an open multimodal assistant.
AI Desk3 min read
Meta's open-weight multimodal mixture-of-experts model, trained for agentic coding, with roughly 816B parameters and a 1M-token context.
Google's family of lightweight open models from 1B to 27B parameters, with image input on larger sizes, 128K context and support for 140+ languages.
OpenAI's contrastive image-text model that embeds images and text in a shared space, enabling zero-shot classification and image search.