v0.22.0

zjunlp/OmniThinkv0.22.0Jul 17, 2026by VikParuchuri

AI Summary

This release introduces a server-based model architecture to improve resource management by sharing models across processes and resolves a critical bug where text detection weights failed to load.

Key Highlights

Switched to a spawn + server model architecture for better resource utilization
Fixed a bug where text detection weights did not load properly
Added support for Vik/fast models

Breaking Changes

Architecture change: Models are now managed via a server process instead of being loaded locally in each process

New Features

Vik/fast models

Full Release Notes

Swap models to a spawn + server model, instead of relying on each process having a separate model.
Fix bug with text detection where weights didn't load properly.

## What's Changed
* Vik/fast models by @VikParuchuri in https://github.com/datalab-to/surya/pull/532
* Dev by @VikParuchuri in https://github.com/datalab-to/surya/pull/533


**Full Changelog**: https://github.com/datalab-to/surya/compare/v0.21.2...v0.22.0