MMOCR Release v1.0.0rc5

Highlights

1. Two models, Aster and SVTR, are added to our model zoo. The full implementation of ABCNet is also available now.
2. Dataset Preparer supports 5 more datasets: CocoTextV2, FUNSD, TextOCR, NAF, SROIE.
3. We have 4 more text recognition transforms, and two helper transforms. See #1646 #1632 #1645 for details.
4. The transform, FixInvalidPolygon, is getting smarter at dealing with invalid polygons, and now capable of handling more weird annotations. As a result, a complete training cycle on TotalText dataset can be performed bug-free. The weights of DBNet and FCENet pretrained on TotalText are also released.

New Features & Enhancements

• Update ic15 det config according to DataPrepare by @Harold-lkk in #1617
• Refactor icdardataset metainfo to lowercase. by @Harold-lkk in #1620
• Add ASTER Encoder by @Mountchicken in #1239
• Add ASTER decoder by @Mountchicken in #1625
• Add ASTER config by @Mountchicken in #1238
• Update ASTER config by @Mountchicken in #1629
• Support browse_dataset.py to visualize original dataset by @xinke-wang in #1503
• Add CocoTextv2 to dataset preparer by @xinke-wang in #1514
• Add Funsd to dataset preparer by @xinke-wang in #1550
• Add TextOCR to Dataset Preparer by @xinke-wang in #1543
• Refine example projects and readme by @gaotongxiao in #1628
• Enhance FixInvalidPolygon, add RemoveIgnored transform by @gaotongxiao in #1632
• ConditionApply by @Harold-lkk in #1646
• Add NAF to dataset preparer by @Mountchicken in #1609
• Add SROIE to dataset preparer by @FerryHuang in #1639
• Add svtr decoder by @willpat1213 in #1448
• Add missing unit tests by @Mountchicken in #1651
• Add svtr encoder by @willpat1213 in #1483
• ABCNet train by @Harold-lkk in #1610
• Totaltext cfgs for DB and FCE by @gaotongxiao in #1633
• Add Aliases to models by @gaotongxiao in #1611
• SVTR transforms by @gaotongxiao in #1645
• Add SVTR framework and configs by @gaotongxiao in #1621
• Issue Template by @Harold-lkk in #1663

Docs

• Add Chinese translation for browse_dataset.py by @xinke-wang in #1647
• updata abcnet doc by @Harold-lkk in #1658
• update the dbnetpp`s readme file by @zhuyue66 in #1626

Bug Fixes

• nn.SmoothL1Loss beta can not be zero in PyTorch 1.13 version by @Harold-lkk in #1616
• ctc loss bug if target is empty by @Harold-lkk in #1618
• Add torch 1.13 by @gaotongxiao in #1619
• Remove outdated tutorial link by @gaotongxiao in #1627
• Dev 1.x some doc mistakes by @KevinNuNu in #1630
• Support custom font to visualize some languages (e.g. Korean) by @ProtossDragoon in #1567
• db_module_loss，negative number encountered in sqrt by @KevinNuNu in #1640
• Use int instead of np.int by @gaotongxiao in #1636
• Remove support for py3.6 by @gaotongxiao in #1660

New Contributors

• @zhuyue66 made their first contribution in #1626
• @KevinNuNu made their first contribution in #1630
• @FerryHuang made their first contribution in #1639
• @willpat1213 made their first contribution in #1448

Full Changelog: v1.0.0rc4...v1.0.0rc5