Vision and OCR Notes: Sarvam, Perceptron AI, Reducto and Meta
The common thread is that data pipelines, not model size, decide whether document and video systems work — parsing, token selection and smart caching each cut cost by an order of magnitude. Builders working on OCR, video understanding or evaluation should copy the concrete tricks, like Reducto's parsed-page input and Meta's verdict caching. Merve Noyan's US$3-4 image-labeling pipeline is a ready template for cheap detection training.
Sarvam's document model read 13 trillion tokens of text before seeing a pixel, and four months after launch it is digitizing 35 million pages. Perceptron AI built a model that decides which visual tokens to read, since an hour of video produces roughly a million tokens but models learn from about 2 percent of them. Reducto found that giving models a parsed version of a PDF page lifted benchmark scores on document-based decisions, where frontier models score about 30 percent. Meta keeps most videos out of its three-agent review pipeline by compressing similar frames, caching verdicts on viral videos, and letting creators with strong records skip review on metadata alone.
From the back row: - Well under one percent of the web crawl frontier models learn from is in an Indian language. Sarvam's document model read thirteen trillion tokens of text before it saw a pixel, and four months after launch it is digitizing 35 million pages. - Feed a model an hour of video and roughly a million visual tokens go in. The only ground truth is a transcript or a few labeled frames, so it learns from about two percent of them. Perceptron AI's fix is a model that decides which tokens to read. - A frontier model scores about thirty percent on a data lab's benchmark of decisions made from PDFs. Reducto found that handing models a parsed version of the page lifted their scores and cut their reasoning tokens. - Merve Noyan's pipeline labels images with a big open model, has two smaller models judge the boxes, and trains a detector on the survivors, for three or four dollars. Her coding agent made extra training images by flipping traffic signs left to right, until told not to. - Meta keeps most videos out of its three agent review pipeline. It compresses similar frames, caches verdicts on viral videos, and lets creators with a strong record skip it on metadata alone. Vision & OCR playlist: youtube.com/watch?v=RQi7x-… 💬 1 🔄 0 ❤️ 2 👀 390 📊 2 ⚡
