Search Results for "ocr"

DeepSeek-OCR is an open-source optical character recognition solution built as part of the broader DeepSeek AI vision-language ecosystem. It is designed to extract text from images, PDFs, and scanned documents, and integrates with multimodal capabilities that understand layout, context, and visual elements beyond raw character recognition. The system treats OCR not simply as “read the text” but as “understand what the text is doing in the image”—for example distinguishing captions from body text, interpreting tables, or recognizing handwritten versus printed words. ...

...Qwen3-VL is built for complex tasks such as GUI automation, multimodal coding (converting images or videos into HTML, CSS, JS, or Draw.io diagrams), long-context reasoning with support up to 1M tokens, and comprehensive video understanding. It also brings advanced perception capabilities, including spatial grounding, object recognition, OCR across 32 languages, and robust handling of challenging inputs like low-light or distorted text.

Qwen3-Omni is a natively end-to-end multilingual omni-modal foundation model that processes text, images, audio, and video and delivers real-time streaming responses in text and natural speech. It uses a Thinker-Talker architecture with a Mixture-of-Experts (MoE) design, early text-first pretraining, and mixed multimodal training to support strong performance across all modalities without sacrificing text or image quality. The model supports 119 text languages, 19 speech input languages, and...

MiniCPM-o 2.6 is a cutting-edge multimodal large language model (MLLM) designed for high-performance tasks across vision, speech, and video. Capable of running on end-side devices such as smartphones and tablets, it provides powerful features like real-time speech conversation, video understanding, and multimodal live streaming. With 8 billion parameters, MiniCPM-o 2.6 surpasses its predecessors in versatility and efficiency, making it one of the most robust models available. It supports...

...The model excels at non-standard layouts and complex table structures, outperforming non-reasoning OCR systems like GPT-4o and OCRFlux, and competing with large closed-source reasoning models like Gemini 2.5. Thinking token usage can range from 20% to 500% of the final answer, depending on task difficulty. NuMarkdown-8B-Thinking is released under the MIT license and supports vLLM and Transformers for deployment.

translategemma-4b-it is a lightweight, state-of-the-art open translation model from Google, built on the Gemma 3 family and optimized for high-quality multilingual translation across 55 languages. It supports both text-to-text translation and image-to-text extraction with translation, enabling workflows such as OCR-style translation of signs, documents, and screenshots. With a compact ~5B parameter footprint and BF16 support, the model is designed to run efficiently on laptops, desktops, and private cloud infrastructure, making advanced translation accessible without heavy hardware requirements. TranslateGemma uses a structured chat template that enforces explicit source and target language codes, ensuring consistent, deterministic behavior and reducing ambiguity in multilingual pipelines. ...