Baidu open-sources Unlimited OCR model based on DeepSeek OCR architecture, 3 billion parameters with only 500 million activated
Baidu open-sourced and launched the Unlimited OCR model on June 22, with a total parameter count of 3 billion, activating only 500 million parameters during inference, aiming to solve the problem that end-to-end OCR models slow down the more they generate when parsing long documents.
Return to original text
A short introduction is retained here; the full text from the source website is not reposted. For specific wording and subsequent updates, please refer to the original article.
Open the original article on ITHome