This repository is the code for the final group project of NYU DS-GA 1008 Deep Learning class, taught by Yann LeCun.
Team member: Jiayi Du, Meiyi He, Ziyu Lei (DHL)
Folders Description:
- starter_files: The files provided by instructors.
- hrnet: Predict binary road map using HRNets (High-resolution Networks).
- yolo: Predict bounding boxes using YOLO network.
- pretrain: AutoEncoder for pretraining on unlabeled data.
- submission: Since it's a competition, we have 3 rounds of submissions.
- visualization: Visualize the predictions made by trained models.
- archived: some legacy code.
Given input images (captured by 6 cameras mounted on ego car):
Predicted Result in the following formats:
