Big commit.

This commit contains all the experiments I done.

configs/retina/retinanet_R-50-FPN_1x_adjustl1.yaml

@@ -0,0 +1,47 @@
+MODEL:
+  META_ARCHITECTURE: "RetinaNet"
+  WEIGHT: "catalog://ImageNetPretrained/MSRA/R-50"
+  RPN_ONLY: True
+  BACKBONE:
+    CONV_BODY: "R-50-FPN"
+    OUT_CHANNELS: 256
+  RPN:
+    USE_FPN: True
+    FG_IOU_THRESHOLD: 0.5
+    BG_IOU_THRESHOLD: 0.4
+    ANCHOR_STRIDE: (4, 8, 16, 32, 64)
+    PRE_NMS_TOP_N_TRAIN: 2000
+    PRE_NMS_TOP_N_TEST: 1000
+    POST_NMS_TOP_N_TEST: 1000
+    FPN_POST_NMS_TOP_N_TEST: 1000
+  ROI_HEADS:
+    USE_FPN: True
+    BATCH_SIZE_PER_IMAGE: 256
+  ROI_BOX_HEAD:
+    POOLER_RESOLUTION: 7
+    POOLER_SCALES: (0.25, 0.125, 0.0625, 0.03125)
+    POOLER_SAMPLING_RATIO: 2
+    FEATURE_EXTRACTOR: "FPN2MLPFeatureExtractor"
+    PREDICTOR: "FPNPredictor"
+DATASETS:
+  TRAIN: ("coco_2017_train",)
+  TEST: ("coco_2017_val",)
+INPUT:
+  MIN_SIZE_TRAIN: (800,)
+  MAX_SIZE_TRAIN: 1333
+  MIN_SIZE_TEST: 800
+  MAX_SIZE_TEST: 1333
+DATALOADER:
+  SIZE_DIVISIBILITY: 32
+SOLVER:
+  # Assume 4 gpus
+  BASE_LR: 0.01
+  WEIGHT_DECAY: 0.0001
+  STEPS: (60000, 80000)
+  MAX_ITER: 90000
+  IMS_PER_BATCH: 16
+RETINANET:
+  RETINANET_ON: True
+  SCALES_PER_OCTAVE: 3
+  STRADDLE_THRESH: -1
+  SELFADJUST_SMOOTH_L1: True

configs/retina/retinanet_R-50-FPN_1x_low_quality_0.2.yaml

@@ -0,0 +1,47 @@
+MODEL:
+  META_ARCHITECTURE: "RetinaNet"
+  WEIGHT: "catalog://ImageNetPretrained/MSRA/R-50"
+  RPN_ONLY: True
+  BACKBONE:
+    CONV_BODY: "R-50-FPN"
+    OUT_CHANNELS: 256
+  RPN:
+    USE_FPN: True
+    FG_IOU_THRESHOLD: 0.5
+    BG_IOU_THRESHOLD: 0.4
+    ANCHOR_STRIDE: (4, 8, 16, 32, 64)
+    PRE_NMS_TOP_N_TRAIN: 2000
+    PRE_NMS_TOP_N_TEST: 1000
+    POST_NMS_TOP_N_TEST: 1000
+    FPN_POST_NMS_TOP_N_TEST: 1000
+  ROI_HEADS:
+    USE_FPN: True
+    BATCH_SIZE_PER_IMAGE: 256
+  ROI_BOX_HEAD:
+    POOLER_RESOLUTION: 7
+    POOLER_SCALES: (0.25, 0.125, 0.0625, 0.03125)
+    POOLER_SAMPLING_RATIO: 2
+    FEATURE_EXTRACTOR: "FPN2MLPFeatureExtractor"
+    PREDICTOR: "FPNPredictor"
+DATASETS:
+  TRAIN: ("coco_2017_train",)
+  TEST: ("coco_2017_val",)
+INPUT:
+  MIN_SIZE_TRAIN: (800,)
+  MAX_SIZE_TRAIN: 1333
+  MIN_SIZE_TEST: 800
+  MAX_SIZE_TEST: 1333
+DATALOADER:
+  SIZE_DIVISIBILITY: 32
+SOLVER:
+  # Assume 4 gpus
+  BASE_LR: 0.01
+  WEIGHT_DECAY: 0.0001
+  STEPS: (60000, 80000)
+  MAX_ITER: 90000
+  IMS_PER_BATCH: 16
+RETINANET:
+  RETINANET_ON: True
+  SCALES_PER_OCTAVE: 3
+  STRADDLE_THRESH: -1
+  LOW_QUALITY_THRESHOLD: 0.4

configs/retina/retinanet_R-50-FPN_1x_low_quality_0.3.yaml

@@ -0,0 +1,47 @@
+MODEL:
+  META_ARCHITECTURE: "RetinaNet"
+  WEIGHT: "catalog://ImageNetPretrained/MSRA/R-50"
+  RPN_ONLY: True
+  BACKBONE:
+    CONV_BODY: "R-50-FPN"
+    OUT_CHANNELS: 256
+  RPN:
+    USE_FPN: True
+    FG_IOU_THRESHOLD: 0.5
+    BG_IOU_THRESHOLD: 0.4
+    ANCHOR_STRIDE: (4, 8, 16, 32, 64)
+    PRE_NMS_TOP_N_TRAIN: 2000
+    PRE_NMS_TOP_N_TEST: 1000
+    POST_NMS_TOP_N_TEST: 1000
+    FPN_POST_NMS_TOP_N_TEST: 1000
+  ROI_HEADS:
+    USE_FPN: True
+    BATCH_SIZE_PER_IMAGE: 256
+  ROI_BOX_HEAD:
+    POOLER_RESOLUTION: 7
+    POOLER_SCALES: (0.25, 0.125, 0.0625, 0.03125)
+    POOLER_SAMPLING_RATIO: 2
+    FEATURE_EXTRACTOR: "FPN2MLPFeatureExtractor"
+    PREDICTOR: "FPNPredictor"
+DATASETS:
+  TRAIN: ("coco_2017_train",)
+  TEST: ("coco_2017_val",)
+INPUT:
+  MIN_SIZE_TRAIN: (800,)
+  MAX_SIZE_TRAIN: 1333
+  MIN_SIZE_TEST: 800
+  MAX_SIZE_TEST: 1333
+DATALOADER:
+  SIZE_DIVISIBILITY: 32
+SOLVER:
+  # Assume 4 gpus
+  BASE_LR: 0.01
+  WEIGHT_DECAY: 0.0001
+  STEPS: (60000, 80000)
+  MAX_ITER: 90000
+  IMS_PER_BATCH: 16
+RETINANET:
+  RETINANET_ON: True
+  SCALES_PER_OCTAVE: 3
+  STRADDLE_THRESH: -1
+  LOW_QUALITY_THRESHOLD: 0.4

configs/retina/retinanet_R-50-FPN_1x_low_quality_0.4.yaml

@@ -0,0 +1,47 @@
+MODEL:
+  META_ARCHITECTURE: "RetinaNet"
+  WEIGHT: "catalog://ImageNetPretrained/MSRA/R-50"
+  RPN_ONLY: True
+  BACKBONE:
+    CONV_BODY: "R-50-FPN"
+    OUT_CHANNELS: 256
+  RPN:
+    USE_FPN: True
+    FG_IOU_THRESHOLD: 0.5
+    BG_IOU_THRESHOLD: 0.4
+    ANCHOR_STRIDE: (4, 8, 16, 32, 64)
+    PRE_NMS_TOP_N_TRAIN: 2000
+    PRE_NMS_TOP_N_TEST: 1000
+    POST_NMS_TOP_N_TEST: 1000
+    FPN_POST_NMS_TOP_N_TEST: 1000
+  ROI_HEADS:
+    USE_FPN: True
+    BATCH_SIZE_PER_IMAGE: 256
+  ROI_BOX_HEAD:
+    POOLER_RESOLUTION: 7
+    POOLER_SCALES: (0.25, 0.125, 0.0625, 0.03125)
+    POOLER_SAMPLING_RATIO: 2
+    FEATURE_EXTRACTOR: "FPN2MLPFeatureExtractor"
+    PREDICTOR: "FPNPredictor"
+DATASETS:
+  TRAIN: ("coco_2017_train",)
+  TEST: ("coco_2017_val",)
+INPUT:
+  MIN_SIZE_TRAIN: (800,)
+  MAX_SIZE_TRAIN: 1333
+  MIN_SIZE_TEST: 800
+  MAX_SIZE_TEST: 1333
+DATALOADER:
+  SIZE_DIVISIBILITY: 32
+SOLVER:
+  # Assume 4 gpus
+  BASE_LR: 0.01
+  WEIGHT_DECAY: 0.0001
+  STEPS: (60000, 80000)
+  MAX_ITER: 90000
+  IMS_PER_BATCH: 16
+RETINANET:
+  RETINANET_ON: True
+  SCALES_PER_OCTAVE: 3
+  STRADDLE_THRESH: -1
+  LOW_QUALITY_THRESHOLD: 0.4

configs/retina/retinanet_R-50-FPN_1x_no_low_quality.yaml

@@ -0,0 +1,47 @@
+MODEL:
+  META_ARCHITECTURE: "RetinaNet"
+  WEIGHT: "catalog://ImageNetPretrained/MSRA/R-50"
+  RPN_ONLY: True
+  BACKBONE:
+    CONV_BODY: "R-50-FPN"
+    OUT_CHANNELS: 256
+  RPN:
+    USE_FPN: True
+    FG_IOU_THRESHOLD: 0.5
+    BG_IOU_THRESHOLD: 0.4
+    ANCHOR_STRIDE: (4, 8, 16, 32, 64)
+    PRE_NMS_TOP_N_TRAIN: 2000
+    PRE_NMS_TOP_N_TEST: 1000
+    POST_NMS_TOP_N_TEST: 1000
+    FPN_POST_NMS_TOP_N_TEST: 1000
+  ROI_HEADS:
+    USE_FPN: True
+    BATCH_SIZE_PER_IMAGE: 256
+  ROI_BOX_HEAD:
+    POOLER_RESOLUTION: 7
+    POOLER_SCALES: (0.25, 0.125, 0.0625, 0.03125)
+    POOLER_SAMPLING_RATIO: 2
+    FEATURE_EXTRACTOR: "FPN2MLPFeatureExtractor"
+    PREDICTOR: "FPNPredictor"
+DATASETS:
+  TRAIN: ("coco_2017_train",)
+  TEST: ("coco_2017_val",)
+INPUT:
+  MIN_SIZE_TRAIN: (800,)
+  MAX_SIZE_TRAIN: 1333
+  MIN_SIZE_TEST: 800
+  MAX_SIZE_TEST: 1333
+DATALOADER:
+  SIZE_DIVISIBILITY: 32
+SOLVER:
+  # Assume 4 gpus
+  BASE_LR: 0.01
+  WEIGHT_DECAY: 0.0001
+  STEPS: (60000, 80000)
+  MAX_ITER: 90000
+  IMS_PER_BATCH: 16
+RETINANET:
+  RETINANET_ON: True
+  SCALES_PER_OCTAVE: 3
+  STRADDLE_THRESH: -1
+  LOW_QUALITY_MATCHES: False

configs/retina/retinanet_R-50-FPN_1x_no_low_quality_adjustl1.yaml

@@ -0,0 +1,48 @@
+MODEL:
+  META_ARCHITECTURE: "RetinaNet"
+  WEIGHT: "catalog://ImageNetPretrained/MSRA/R-50"
+  RPN_ONLY: True
+  BACKBONE:
+    CONV_BODY: "R-50-FPN"
+    OUT_CHANNELS: 256
+  RPN:
+    USE_FPN: True
+    FG_IOU_THRESHOLD: 0.5
+    BG_IOU_THRESHOLD: 0.4
+    ANCHOR_STRIDE: (4, 8, 16, 32, 64)
+    PRE_NMS_TOP_N_TRAIN: 2000
+    PRE_NMS_TOP_N_TEST: 1000
+    POST_NMS_TOP_N_TEST: 1000
+    FPN_POST_NMS_TOP_N_TEST: 1000
+  ROI_HEADS:
+    USE_FPN: True
+    BATCH_SIZE_PER_IMAGE: 256
+  ROI_BOX_HEAD:
+    POOLER_RESOLUTION: 7
+    POOLER_SCALES: (0.25, 0.125, 0.0625, 0.03125)
+    POOLER_SAMPLING_RATIO: 2
+    FEATURE_EXTRACTOR: "FPN2MLPFeatureExtractor"
+    PREDICTOR: "FPNPredictor"
+DATASETS:
+  TRAIN: ("coco_2017_train",)
+  TEST: ("coco_2017_val",)
+INPUT:
+  MIN_SIZE_TRAIN: (800,)
+  MAX_SIZE_TRAIN: 1333
+  MIN_SIZE_TEST: 800
+  MAX_SIZE_TEST: 1333
+DATALOADER:
+  SIZE_DIVISIBILITY: 32
+SOLVER:
+  # Assume 4 gpus
+  BASE_LR: 0.01
+  WEIGHT_DECAY: 0.0001
+  STEPS: (60000, 80000)
+  MAX_ITER: 90000
+  IMS_PER_BATCH: 16
+RETINANET:
+  RETINANET_ON: True
+  SCALES_PER_OCTAVE: 3
+  STRADDLE_THRESH: -1
+  LOW_QUALITY_MATCHES: False
+  SELFADJUST_SMOOTH_L1: True

configs/retina/retinanet_mask_R-50-FPN_1.5x.yaml

@@ -0,0 +1,58 @@
+MODEL:
+  META_ARCHITECTURE: "RetinaNet"
+  WEIGHT: "catalog://ImageNetPretrained/MSRA/R-50"
+  RPN_ONLY: True
+  BACKBONE:
+    CONV_BODY: "R-50-FPN"
+    OUT_CHANNELS: 256
+  RPN:
+    USE_FPN: True
+    FG_IOU_THRESHOLD: 0.5
+    BG_IOU_THRESHOLD: 0.4
+    ANCHOR_STRIDE: (4, 8, 16, 32, 64)
+    PRE_NMS_TOP_N_TRAIN: 2000
+    PRE_NMS_TOP_N_TEST: 1000
+    POST_NMS_TOP_N_TEST: 1000
+    FPN_POST_NMS_TOP_N_TEST: 1000
+  ROI_HEADS:
+    USE_FPN: True
+    BATCH_SIZE_PER_IMAGE: 256
+  ROI_BOX_HEAD:
+    POOLER_RESOLUTION: 7
+    POOLER_SCALES: (0.25, 0.125, 0.0625, 0.03125)
+    POOLER_SAMPLING_RATIO: 2
+    FEATURE_EXTRACTOR: "FPN2MLPFeatureExtractor"
+    PREDICTOR: "FPNPredictor"
+  ROI_MASK_HEAD:
+    POOLER_SCALES: (0.125, 0.0625, 0.03125)
+    #POOLER_SCALES: (0.25, 0.125, 0.0625, 0.03125)
+    FEATURE_EXTRACTOR: "MaskRCNNFPNFeatureExtractor"
+    PREDICTOR: "MaskRCNNC4Predictor"
+    POOLER_RESOLUTION: 14
+    POOLER_SAMPLING_RATIO: 2
+    RESOLUTION: 28
+    SHARE_BOX_FEATURE_EXTRACTOR: False
+  MASK_ON: True
+DATASETS:
+  TRAIN: ("coco_2017_train",)
+  TEST: ("coco_2017_val",)
+INPUT:
+  MIN_SIZE_TRAIN: (800,)
+  MAX_SIZE_TRAIN: 1333
+  MIN_SIZE_TEST: 800
+  MAX_SIZE_TEST: 1333
+DATALOADER:
+  SIZE_DIVISIBILITY: 32
+SOLVER:
+  # Assume 4 gpus
+  BASE_LR: 0.005
+  WEIGHT_DECAY: 0.0001
+  STEPS: (180000, 240000)
+  MAX_ITER: 270000
+  IMS_PER_BATCH: 8
+RETINANET:
+  RETINANET_ON: True
+  BACKBONE: "p3p7"
+  SCALES_PER_OCTAVE: 3
+  STRADDLE_THRESH: -1
+  NUM_MASKS_TEST: 50