Open h123a456rr789y opened 2 years ago
Thank you for your interest in our work.
In theory, there is such a risk. However, some datasets do not provide validation datasets at all. Meanwhile, the effect claimed in our paper can be achieved without complex parameter adjustment in practice, so we ignore this.
If you're interested in this, you can use our public code to verify it. Evaluating on validation sets won't significantly reduce the effectiveness of our approach.
Hi, thanks for sharing the great work! I have notice that you have used split 2 for both validation and testing in your code of main_h36m_3d.py and main_3dpw_3d.py I doubt if it's unfair to use the same data for testing and validation
Thanks