whats up with all the recent OCR models being released
#30
by evewashere - opened
Nanonets, MinerU2.5, Deepseek...
What does it all mean?
It might just be a coincidence, or perhaps as more real-world applications based on large models emerge, OCR is gradually becoming a bottleneck for deployment. In any case, PaddleOCR has been deeply cultivating this field for many years. For us, releasing this model is simply a routine update following our development work.