training chinese dataset

Media-Smart / vedastr

A scene text recognition toolbox based on PyTorch

Apache License 2.0

535 stars 100 forks source link

Closed lits901 closed 3 years ago

lits901 commented 3 years ago

how to train chinese dataset? there should be several points need to be modified, right?

Single430 commented 3 years ago

@ChaseMonsterAway @hxcai @mileistone 请问有交流群吗？是否可以在好未来等中文数据集上做一些测试？非常感谢！

ChaseMonsterAway commented 3 years ago

how to train chinese dataset? there should be several points need to be modified, right?

There're some researchers used vedastr do experiments on chinese dataset. Maybe you should change the character set in config file.

ChaseMonsterAway commented 3 years ago

@ChaseMonsterAway @hxcai @mileistone 请问有交流群吗？是否可以在好未来等中文数据集上做一些测试？非常感谢！

@ChaseMonsterAway @hxcai @mileistone 请问有交流群吗？是否可以在好未来等中文数据集上做一些测试？非常感谢！

You can refer to #44 for communication group. Currently, we have no plan to train a model for chinese. You can try and pull a request!