How to train with my own dataset

FangGet / PackNet-SFM-PyTorch

A Pytorch implementation(unofficial) for paper "PackNet-SfM: 3D Packing for Self-Supervised Monocular Depth Estimation"

GNU General Public License v3.0

21 stars 7 forks source link

How to train with my own dataset #9

Open wangsigong opened 2 years ago

wangsigong commented 2 years ago

How do I train with my own dataset? Can I complete my training directly using the continuous monocular video I collected? I don't have radar, speed, etc., what specific training data does he need?