_BASE_: ./maskformer_R50_bs16_160k.yaml
MODEL:
  BACKBONE:
    NAME: "D2ViTTransformer"
    FREEZE_AT: -1
  SWIN:
    EMBED_DIM: 768
    DEPTHS: [2, 2, 6, 2]
    NUM_HEADS: [3, 6, 12, 24]
    WINDOW_SIZE: 7
    APE: False
    DROP_PATH_RATE: 0.3
    PATCH_NORM: True
  PIXEL_MEAN: [123.675, 116.280, 103.530]
  PIXEL_STD: [58.395, 57.120, 57.375]
  WEIGHTS: None
  MASK_FORMER:
    NUM_OBJECT_QUERIES: 4
  SEM_SEG_HEAD:
    PIXEL_DECODER_NAME: BigPixelDecoder
SOLVER:
  BASE_LR: 0.00015
  IMS_PER_BATCH: 8
  WARMUP_FACTOR: 1e-6
  WARMUP_ITERS: 1500
  WEIGHT_DECAY: 0.01
  WEIGHT_DECAY_NORM: 0.0
  WEIGHT_DECAY_EMBED: 0.0
  BACKBONE_MULTIPLIER: 1.0
FLAGS:
  UNFREEZE_AT: []