ade20k_segmenter.yaml

DEVICE          : cuda              # device used for training and evaluation (cpu, cuda, cuda0, cuda1, ...)
SAVE_DIR        : 'output-dir-path'         # output folder name used for saving the model, logs and inference results
ADDENDUM        : 'will-be-appended-at-the-end-of-logfoldername'

MODEL:                                    
  NAME          : SegMenter                                           # name of the model you are using
  BACKBONE      : vit_small_patch16_224                                                  # model variant
  PRETRAINED    : 'Location-of-pretrained-backbone-model' 

DATASET:
  NAME          : ADE20K                                              # dataset name to be trained with (camvid, cityscapes, ade20k)
  ROOT          : data-dir-path                         # dataset root path
  IGNORE_LABEL  : -1
  N_CLS         : 151
  SEED          : 0
  
TRAIN:
  BASE_SIZE     : 520
  IMAGE_SIZE    : [512, 512]      # training image size in (h, w)
  BATCH_SIZE    : 14                 # batch size used to train
  EPOCHS        : 32             # number of epochs to train
  EVAL_INTERVAL : 32              # evaluation interval during training
  ADVERSARIAL   : true
  ATTACK        : pgd
  LOSS_FN       : pgd
  EPS           : 4
  N_ITERS       : 5
  FREEZE        : false
  AMP           : false           # use AMP in training
  DDP           : true           # use DDP training

LOSS:
  NAME          : CrossEntropy          # loss function name (ohemce, ce, dice)
  CLS_WEIGHTS   : false            # use class weights in loss calculation

OPTIMIZER:
  NAME          : sgd           # optimizer name, default=sgd
  LR            : 0.002          # initial learning rate used in optimizer
  WEIGHT_DECAY  : 0.01            # decay rate used in optimizer 

SCHEDULER:
  NAME          : polylr    # scheduler name
  POWER         : 0.9             # scheduler power
  WARMUP        : 5              # warmup epochs used in scheduler
  WARMUP_RATIO  : 0.00001             # warmup ratio
  

EVAL:
  NAME          : ADE20K                                              # dataset name to be trained with (camvid, cityscapes, ade20k)
  BACKBONE      : vit_small_patch16_224                                                 # model variant
  N_CLS         : 151
  MODEL_PATH    : 'path to checkpoint of evaluation model'
  BASE_SIZE     : 520
  IMAGE_SIZE    : [512, 512]        
  BATCH_SIZE    : 24