v0.4.15

datalab-to/suryav0.4.15Jul 12, 2024by VikParuchuri

AI Summary

Switches text detection and layout model architectures to significantly boost performance across different hardware accelerators.

Key Highlights

  • 30% faster on GPU
  • 4x faster on CPU
  • 12x faster on MPS (Mac)
  • Maintained or slightly improved accuracy

New Features

  • New text detection and layout architecture

Full Release Notes

Switched model architecture for the text detection and layout models:

- 30% faster on GPU
- 4x faster on CPU
- 12x faster on MPS (M series macs)

Accuracy should be about the same, or slightly better, from my benchmarks.