search

facebookresearch / DetectAndTrack

The implementation of an algorithm presented in the CVPR18 paper: "Detect-and-Track: Efficient Pose Estimation in Videos"
Apache License 2.0
1k stars 186 forks source link

Regarding training mode. #28

Closed imnaren142 closed 6 years ago

imnaren142 commented 6 years ago

Hi, how to use the downloaded data to train the model? For training we don't have the "PoseTrackV1.0_Annots_val_json" folder?? Can you share the steps to reproduce the training part?

StalonWilson commented 6 years ago

@naren142 You may try downloading these files from this website. Then, extract them and put them into a new folder lib/datasets/data/PoseTrackV1.0_Annots_val_json which created by you.

imnaren142 commented 6 years ago

@StalonWilson Hey i am talking about "ANN_DN" in the json_dataset.py file.What should be the value there for training?? For validation part it is the folder "PoseTrackV1.0_Annots_val_json" right?

rohitgirdhar commented 6 years ago

As mentioned in the README, this folder is only required for evaluation. It can have any arbitrary value while training.