🚀 Features
- add use_preprocess_img and use_vertical_padding logic by @SWHL in 3cc33f8
- add use_preprocess_img and use_apply_vertical_padding parameter by @SWHL in 83866d6
- verify RapidOCR model accuracy under TensorRT inference engine. Details in blog: RapidOCR PP-OCRv6 Supports TensorRT Inference Engine
🐛 Bug Fixes
- improve IDE support for lazy package exports by @SWHL in 0962933
- fixed issue #695 by @SWHL in 0fa7a96
- fixed issue #709 by @SWHL in 802c056
- align the performance of the torch version of the PP-OCRv6 Rec tiny model by @SWHL in b64a46f
🧪 Testing
- remove invalid unit testings of tensorrt module by @SWHL in c4db82f
- add PP-OCRv3 mobile unit testings by @SWHL in bce0c99
🎉 Contributors
Full Changelog: 3.9.2