Open HUAFOR opened 1 year ago
Hi, thanks for your interest. Our evaluation code is based on this repo. Please refer to it and our paper for more details. Thanks~
非常感谢您的回复!我想请问您是否可以考虑提供 测试得到上述表格(Table1 in the paper)各列指标数据 所用到的源码(例如:inference_multi.py)?对于我来说复现这部分的逻辑似乎有些困难。
(例如:inference_multi.py)?对于我来说复现这部分的逻辑似乎有些困难。
Hi, sorry, every day we have so many things to do. We have tried our best to rearrange the necessary code for readability and release it.
It should be easy to get the code you want from inference_single.py
When comparing with other benchmark models, do we need to ensure that all models are trained on the same dataset? Thanks!
非常感谢作者您为社区带来的开源贡献!我想知道如何通过您提供的checkpoint来复现您论文中表格1里面的精度数据(以lrs2为例) inference_single.py只提供了一段测试输出的视频,没有提供相应的精度数据~