Tag

multimodal-ai.

2 writings found

Latest Archives

NeoMME: The Efficient Multimodal Encoder That Skips the VLM Bloat

NeoMME ditches separate vision towers for unified multimodal encoding. I break down why this matters for document retrieval and what it means for building efficient AI systems.

NeoMME: Building Document Retrieval Without Pretrained VLM Overhead

NeoMME is a 260M/800M multimodal encoder that rethinks visual document retrieval by training a single bidirectional Transformer from scratch instead of stacking pretrained components.
View all rollups →