lars_val_semantic_lbl.yaml

MODE: 'semantic'

PATHS:
  # Val dataset root
  DATASET_ROOT: 'path/to/LaRS/val'
  # Path to the predictions root dir (contains prediction dirs for each method)
  PREDICTIONS: '/path/to/predictions/semantic'
  # Evaluation results root output dir
  RESULTS: './results/v1.0.0/semantic'

DATASET:
  IMAGE_SUBDIR: 'images'
  SEMANTIC_MASK_SUBDIR: 'semantic_masks'
  PANOPTIC_MASK_SUBDIR: 'panoptic_masks'
  PANOPTIC_ANN_FILE: 'panoptic_annotations.json'
  SUBSET_LIST: 'image_list.txt'

SEGMENTATION:
  # Colors used in the predictions masks
  COLORS: 
    - 0 # Obstacles id
    - 1 # Water id
    - 2 # Sky id

PROGRESS:
  METRICS: ['Q', 'mIoU', 'WE_acc', 'F1']