GLEE/projects/GLEE/configs/videos/Lite/ovis_base.yaml

35 lines
770 B
YAML

_BASE_: "../../images/Lite/base_clip_frozen_image_R50.yaml"
MODEL:
PSEUDO_VIDEO: False
FREEZE_WHOLE: False
TEXT:
ARCH: clip_frozen
DATASETS:
TRAIN: ("ovis_train",)
TEST: ("ovis_val",)
SOLVER:
IMS_PER_BATCH: 8
BASE_LR: 0.0001
STEPS: (12000, )
MAX_ITER: 18000
CHECKPOINT_PERIOD: 2000
INPUT:
SAMPLING_FRAME_NUM: 2
SAMPLING_FRAME_RANGE: 10
MIN_SIZE_TRAIN_SAMPLING: "choice_by_clip"
RANDOM_FLIP: "flip_by_clip"
MIN_SIZE_TRAIN: (480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800)
# MAX_SIZE_TRAIN: 1333
MIN_SIZE_TEST: 720
CROP:
ENABLED: True
TYPE: "absolute_range"
SIZE: (384, 600)
FORMAT: "RGB"
TEST:
EVAL_PERIOD: 100000
DATALOADER:
FILTER_EMPTY_ANNOTATIONS: False
NUM_WORKERS: 8
OUTPUT_DIR: ./GLEE_Lite_ovis