yamle.third_party.nyuv2 package#

author: Mihai Suteu date: 15/05/19

Adapted from: xapharius/pytorch-nyuv2

class yamle.third_party.nyuv2.NYUv2(root, train=True, download=False, task='segmentation')[source]#

Bases: Dataset

PyTorch wrapper for the NYUv2 dataset. Data sources available: Semantic Segmentation and Depth Estimation.

1. Semantic Segmentation: 1 channel representing one of the 14 (0 - background) classes.

  1. Depth Images: 1 channel with floats.

Parameters:

  • root (str) – Root directory of dataset where NYUv2/processed/training.pt and NYUv2/processed/test.pt exist.

  • train (bool, optional) – If True, creates dataset from training.pt,

  • download (bool, optional) – If true, downloads the dataset from the internet and

  • task (str, optional) – Choose from: segmentation, depth.

download()[source]#
yamle.third_party.nyuv2.download_rgb(root)[source]#
yamle.third_party.nyuv2.download_seg(root)[source]#
yamle.third_party.nyuv2.download_depth(root)[source]#