Key Info

OpenBMB has released MiniCPM-OCR, a 1B-parameter model aimed at document parsing. A Hugging Face Space built by Niels Rogge lets users compare its outputs side by side with Baidu's new Unlimited OCR v2.

Highlights

  • MiniCPM-OCR is a 1B-parameter OCR model focused on document parsing.
  • It comes from OpenBMB.
  • A public Hugging Face Space provides output comparisons against Baidu's Unlimited OCR v2.
  • The comparison tool is intended to help users evaluate the two OCR models directly.