TrOCR Large Printed is Microsoft’s transformer-based optical character recognition model for recognizing printed text in images. Developers and researchers can download and run the model locally through the Hugging Face ecosystem for document and image text extraction.
Trocr Large Printed is a Computer vision, OCR & document AI project. It focuses on extracting machine-readable text from printed document images without building an OCR model from scratch. It is built as an open-source project for machine learning developers and researchers. Trocr Large Printed is open source under the MIT license. It ships for the web and API, and it can be self-hosted.
Microsoft builds and maintains Trocr Large Printed, and it first shipped in 2019. The project is developed in the open on GitHub with 22.2k stars. Key capabilities include image text recognition, printed text OCR, and transformer architecture.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do