中文ocr效果极差
lucasjinreal opened this issue · comments
1.5阶段不是加入了中文ocr数据么，为什么识别中文依旧没有任何效果
The model checkpoint in this blog are NOT released yet. All the public checkpoints are trained in English only at this moment.
https://llava-vl.github.io/blog/2024-05-25-llava-next-ablations/
When will the Chinese weights open?
@ChunyuanLI Once again, catching eyeballs but delaying the release of the code and weights. Disappointing.