A high-performance, privacy-focused OCR solution that runs entirely in the browser using ONNX Runtime with both RapidOCR and PPU PaddleOCR models. Process text from images and PDF documents without ...
Adapter-based optical compression for long documents using DeepSeek's DeepEncoder with Qwen3-VL-2B. Built on DeepSeek-OCR by DeepSeek-AI. This repo includes deepencoder.py from DeepSeek-OCR ...
Betting that the biggest barrier to enterprise AI is the “paper problem,” Mistral AI released its third-generation optical character recognition (OCR) model on Tuesday. Paris-based Mistral claims its ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback