DeepSeek has unveiled DeepSeek-OCR, an open-source AI model that processes text by converting it into images using 2D mapping. This new “Context Optical Compression” approach allows large language models to handle longer documents with fewer tokens and improved accuracy. The method has drawn praise for its potential to simplify and speed up text reasoning.
DeepSeek has unveiled DeepSeek-OCR, an open-source AI model that processes text by converting it into images using 2D mapping. This new “Context Optical Compression” approach allows large language models to handle longer documents with fewer tokens and improved accuracy. The method has drawn praise for its potential to simplify and speed up text reasoning.
