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
687 stars 91 forks source link

rootnet_output #13

Open Ariana8456 opened 4 years ago

Ariana8456 commented 4 years ago

Hi： I can't find this two file rootnet_interhand2.6m_output_machine_annot_val.json and rootnet_interhand2.6m_output_all_test.json in the dataset,please tell me the files where I can find?

mks0601 commented 4 years ago

It is included in the README. https://github.com/facebookresearch/InterHand2.6M/releases/download/v0.0/RootNet.output.on.InterHand2.6M.zip

Ariana8456 commented 4 years ago

Hi,when I run the test.py,the folder of output/result and output/vis are nothing.I can't see the prediction of hand keypoints in the test images.Could you tell me this porblem how to solve? I run your code in the offline docker. and the docker don't support visualization.

mks0601 commented 4 years ago

Please set vis=True at here and here.

Ariana8456 commented 4 years ago

Hi: Your work is great! I also want to see other resnet's result, such as set resnet_type= 'resnext50_32x4d'.So I add a line 5032:(Bottleneck, [3, 4, 6, 3],[64, 256, 512, 1024, 2048], 'resnext50_32x4d') in /common/nets/resnet.py.But it arise a KeyError: 'resnext50_32x4d'. Could you please tell me how to try resnext50_32x4d? 捕获

mks0601 commented 4 years ago

You'd better search torchvision model zoo to know how to use resnext? I use torchvision model zoo to load resnets.