DeepSeek-OCR
DeepSeek-OCR is an open source model for Contexts Optical Compression, built to explore the boundaries of visual-text compression and investigate the role of vision encoders from an LLM-centric viewpoint. It is designed to compress long contexts through optical 2D mapping, using DeepEncoder as the core engine and DeepSeek3B-MoE-A570M as the decoder. DeepEncoder maintains low activations under high-resolution input while achieving high compression ratios, keeping the number of vision tokens manageable for document understanding. The model supports OCR and document parsing workflows for images and PDFs, with inference through vLLM or Transformers. Users can run image OCR with streaming output, process PDFs with high concurrency, or run batch evaluation for benchmarks. DeepSeek-OCR can convert documents to Markdown, perform free OCR without layouts, parse figures, describe images in detail, and locate referenced text inside an image.
Learn more
HunyuanOCR
Tencent Hunyuan is a large-scale, multimodal AI model family developed by Tencent that spans text, image, video, and 3D modalities, designed for general-purpose AI tasks like content generation, visual reasoning, and business automation. Its model lineup includes variants optimized for natural language understanding, multimodal vision-language comprehension (e.g., image & video understanding), text-to-image creation, video generation, and 3D content generation. Hunyuan models leverage a mixture-of-experts architecture and other innovations (like hybrid “mamba-transformer” designs) to deliver strong performance on reasoning, long-context understanding, cross-modal tasks, and efficient inference. For example, the vision-language model Hunyuan-Vision-1.5 supports “thinking-on-image”, enabling deep multimodal understanding and reasoning on images, video frames, diagrams, or spatial data.
Learn more
CodeT5
Code for CodeT5, a new code-aware pre-trained encoder-decoder model. Identifier-aware unified pre-trained encoder-decoder models for code understanding and generation. This is the official PyTorch implementation for the EMNLP 2021 paper from Salesforce Research. CodeT5-large-ntp-py is specially optimized for Python code generation tasks and employed as the foundation model for our CodeRL, yielding new SOTA results on the APPS Python competition-level program synthesis benchmark. This repo provides the code for reproducing the experiments in CodeT5. CodeT5 is a new pre-trained encoder-decoder model for programming languages, which is pre-trained on 8.35M functions in 8 programming languages (Python, Java, JavaScript, PHP, Ruby, Go, C, and C#). In total, it achieves state-of-the-art results on 14 sub-tasks in a code intelligence benchmark - CodeXGLUE. Generate code based on the natural language description.
Learn more
PrecisionOCR
PrecisionOCR is a ready-to-use, secure, HIPAA-compliant, cloud-based platform for extracting medical meaning from unstructured documents using Optical Character Recognition (OCR).
PrecisionOCR uses custom Optical Character Recognition and AI algorithms to convert PDFs/JPEGs/PNGs into structured, searchable documents. Organizations can work with our team to build OCR report extractors which look for specific types of information to extract or highlight to reduce the noise that comes from extracting all of the data within a document.
Natural language processing (NLP) and machine learning (ML) power the semi-automated and automated transformation of source material such as pdfs or images into structured data records that integrate seamlessly with EMR data using HL7s FHIR standards. Data can be automatically stored along side patient records.
Our OCR document classification is also available along with multiple ways to integrate including API and CLI support.
Learn more