search

facebookresearch / InterHand2.6M

Official PyTorch implementation of "InterHand2.6M: A Dataset and Baseline for 3D Interacting Hand Pose Estimation from a Single RGB Image", ECCV 2020
Other
676 stars 92 forks source link

Question about InterNet compared with other work #34

Open shawnricecake opened 3 years ago

shawnricecake commented 3 years ago

Hi,

I read the paper (ECCV), and I found that you compared your InterNet with other state of art work in Table 5.

So, I would like to know how can I get the EPE of the model on dataset STB and RHP?

By the way, the model you used to get EPE on STB or RHP is trained by InterHand dataset or the corresponding dataset? (I mean the corresponding is that if we want to get EPE on STB data, we need to train the model on STB dataset)

mks0601 commented 3 years ago

You can train InterNet on STB and RHP dataset. The results of the table are from training InterNet on corresponding datasets.

shawnricecake commented 3 years ago

You can train InterNet on STB and RHP dataset. The results of the table are from training InterNet on corresponding datasets.

Yeah, and how to test the model on STB or RHP to get EPE?

mks0601 commented 3 years ago

Set dataset of config.py to STB or RHD and run python test.py --gpu 0-3 --test_epoch $EPOCH.

shawnricecake commented 3 years ago

Set dataset of config.py to STB or RHD and run python test.py --gpu 0-3 --test_epoch $EPOCH.

OK, thanks a lot, just revise the config.py and run the code similar to the InterHand.

mks0601 commented 3 years ago

Yeap