Search Results for "benchmark"

...It is designed to handle complex layouts and noisy documents by giving the model causal reasoning capabilities that mimic human visual scanning behavior, enhancing OCR performance on documents with rich spatial structure. The repository provides model code and inference scripts that let researchers and developers run and benchmark the system on both images and PDFs, with support for batch evaluation and optimized pipelines leveraging vLLM and transformers.

OpenOCR is an open-source General OCR toolkit developed by the OCR team at Fudan University for research and real-world document processing applications. It provides a unified platform for text detection, text recognition, formula recognition, table recognition, and document parsing. Built on advanced OCR technologies such as SVTRv2 and UniRec-0.1B, OpenOCR delivers high accuracy while maintaining efficient inference performance. The toolkit supports both Chinese and English content, making...

...Despite being fairly lightweight (about 1 billion parameters), it delivers state-of-the-art performance across a wide variety of OCR tasks, outperforming many traditional OCR systems and even other multimodal models on benchmark suites. HunyuanOCR handles complex documents: multi-column layouts, tables, mathematical formulas, mixed languages, handwritten or stylized fonts, receipts, tickets, and even video-frame subtitles. The project provides code, pretrained weights, and inference instructions, making it feasible to deploy locally or on a server, and to integrate with applications.

...You can send the input PDFs (by file name or by byte array) and you can have the final PDF (saved on a file or get back on a byte array). The library calls can be synchronous or asynchronous. We want to give you a benchmark, the library was used to create a PDF from single page(scanned) image by an OCR SDK (it is not included in our library, you can use any on the market): 20,000 Images (the OCR SDK creates single page PDF text searchable, running 50 threads) in 80 minutes. The size of the final PDF searchable was 800Mb. If you download the library, we provide a sample which cover all the scenarios possible (synchronous and asynchronous).