official pytorch code

.gitignore

@@ -127,3 +127,7 @@ dmypy.json
 
 # Pyre type checker
 .pyre/
+
+
+checkpoints/
+logs/

DeepLab-V2-PyTorch/configs/coco.yaml

@@ -0,0 +1,58 @@
+EXP:
+    ID: coco
+    OUTPUT_DIR: data
+
+DATASET:
+    NAME: coco
+    ROOT:
+    LABELS: ./data/datasets/coco/labels.txt
+    N_CLASSES: 91
+    IGNORE_LABEL:
+    SCALES:
+    SPLIT:
+        TRAIN:
+        VAL:
+        TEST:
+
+DATALOADER:
+    NUM_WORKERS: 0
+
+IMAGE:
+    MEAN:
+        R: 122.675
+        G: 116.669
+        B: 104.008
+    SIZE:
+        BASE:
+        TRAIN:
+        TEST: 513
+
+MODEL:
+    NAME: DeepLabV1_ResNet101
+    N_BLOCKS: [3, 4, 23, 3]
+    ATROUS_RATES:
+    INIT_MODEL:
+
+SOLVER:
+    BATCH_SIZE:
+        TRAIN: 5
+        TEST: 1
+    ITER_MAX: 100000
+    ITER_SIZE: 2
+    ITER_SAVE: 5000
+    ITER_TB: 100
+    LR_DECAY: 10
+    LR: 2.5e-4
+    MOMENTUM: 0.9
+    OPTIMIZER: sgd
+    POLY_POWER: 0.9
+    WEIGHT_DECAY: 5.0e-4
+    AVERAGE_LOSS: 20
+
+CRF:
+    ITER_MAX: 10
+    POS_W: 3
+    POS_XY_STD: 1
+    BI_W: 4
+    BI_XY_STD: 67
+    BI_RGB_STD: 3

DeepLab-V2-PyTorch/configs/cocostuff10k.yaml

@@ -0,0 +1,58 @@
+EXP:
+    ID: cocostuff10k
+    OUTPUT_DIR: data
+
+DATASET:
+    NAME: cocostuff10k
+    ROOT: /media/kazuto1011/Extra/cocostuff/cocostuff-10k-v1.1
+    LABELS: ./data/datasets/cocostuff/labels.txt
+    N_CLASSES: 182
+    IGNORE_LABEL: 255
+    SCALES: [0.5, 0.75, 1.0, 1.25, 1.5]
+    SPLIT:
+        TRAIN: train
+        VAL: test
+        TEST:
+
+DATALOADER:
+    NUM_WORKERS: 0
+
+IMAGE:
+    MEAN:
+        R: 122.675
+        G: 116.669
+        B: 104.008
+    SIZE:
+        BASE:
+        TRAIN: 321
+        TEST: 513
+
+MODEL:
+    NAME: DeepLabV2_ResNet101_MSC
+    N_BLOCKS: [3, 4, 23, 3]
+    ATROUS_RATES: [6, 12, 18, 24]
+    INIT_MODEL: data/models/coco/deeplabv1_resnet101/caffemodel/deeplabv1_resnet101-coco.pth
+
+SOLVER:
+    BATCH_SIZE:
+        TRAIN: 5
+        TEST: 5
+    ITER_MAX: 20000
+    ITER_SIZE: 2
+    ITER_SAVE: 5000
+    ITER_TB: 100
+    LR_DECAY: 10
+    LR: 2.5e-4
+    MOMENTUM: 0.9
+    OPTIMIZER: sgd
+    POLY_POWER: 0.9
+    WEIGHT_DECAY: 5.0e-4
+    AVERAGE_LOSS: 20
+
+CRF:
+    ITER_MAX: 10
+    POS_W: 3
+    POS_XY_STD: 1
+    BI_W: 4
+    BI_XY_STD: 67
+    BI_RGB_STD: 3

DeepLab-V2-PyTorch/configs/cocostuff164k.yaml

@@ -0,0 +1,58 @@
+EXP:
+    ID: cocostuff164k
+    OUTPUT_DIR: data
+
+DATASET:
+    NAME: cocostuff164k
+    ROOT: /media/kazuto1011/Extra/cocostuff/cocostuff-164k
+    LABELS: ./data/datasets/cocostuff/labels.txt
+    N_CLASSES: 182
+    IGNORE_LABEL: 255
+    SCALES: [0.5, 0.75, 1.0, 1.25, 1.5]
+    SPLIT:
+        TRAIN: train2017
+        VAL: val2017
+        TEST:
+
+DATALOADER:
+    NUM_WORKERS: 0
+
+IMAGE:
+    MEAN:
+        R: 122.675
+        G: 116.669
+        B: 104.008
+    SIZE:
+        BASE: # None
+        TRAIN: 321
+        TEST: 513
+
+MODEL:
+    NAME: DeepLabV2_ResNet101_MSC
+    N_BLOCKS: [3, 4, 23, 3]
+    ATROUS_RATES: [6, 12, 18, 24]
+    INIT_MODEL: data/models/coco/deeplabv1_resnet101/caffemodel/deeplabv1_resnet101-coco.pth
+
+SOLVER:
+    BATCH_SIZE:
+        TRAIN: 5
+        TEST: 1
+    ITER_MAX: 100000
+    ITER_SIZE: 2
+    ITER_SAVE: 5000
+    ITER_TB: 100
+    LR_DECAY: 10
+    LR: 2.5e-4
+    MOMENTUM: 0.9
+    OPTIMIZER: sgd
+    POLY_POWER: 0.9
+    WEIGHT_DECAY: 5.0e-4
+    AVERAGE_LOSS: 20
+
+CRF:
+    ITER_MAX: 10
+    POS_W: 3
+    POS_XY_STD: 1
+    BI_W: 4
+    BI_XY_STD: 67
+    BI_RGB_STD: 3

DeepLab-V2-PyTorch/configs/conda_env.yaml

@@ -0,0 +1,26 @@
+name: deeplab-pytorch
+channels:
+  - pytorch
+  - conda-forge
+  - defaults
+dependencies:
+  # - clang # For MacOS
+  # - clangxx # For MacOS
+  - click
+  - cudatoolkit=10.0
+  - cython
+  - matplotlib
+  - pydensecrf
+  - pytorch
+  - pyyaml
+  - scipy
+  - torchvision
+  - tqdm
+  - pip:
+    - addict
+    - black
+    - joblib
+    - opencv-python
+    - tensorboardX
+    - tensorflow
+    - torchnet

DeepLab-V2-PyTorch/configs/voc12.yaml

@@ -0,0 +1,58 @@
+EXP:
+    ID: voc12
+    OUTPUT_DIR: data
+
+DATASET:
+    NAME: vocaug
+    ROOT: /data/DB/
+    LABELS: ./data/datasets/voc12/labels.txt
+    N_CLASSES: 21
+    IGNORE_LABEL: 255
+    SCALES: [0.5, 0.75, 1.0, 1.25, 1.5]
+    SPLIT:
+        TRAIN: train_cls
+        VAL: val_cls
+        TEST: test
+
+DATALOADER:
+    NUM_WORKERS: 0
+
+IMAGE:
+    MEAN:
+        R: 122.675
+        G: 116.669
+        B: 104.008
+    SIZE:
+        BASE: # None
+        TRAIN: 321
+        TEST: 513
+
+MODEL:
+    NAME: DeepLabV2_ResNet101_MSC
+    N_BLOCKS: [3, 4, 23, 3]
+    ATROUS_RATES: [6, 12, 18, 24]
+    INIT_MODEL: data/models/coco/deeplabv1_resnet101/caffemodel/deeplabv1_resnet101-coco.pth
+
+SOLVER:
+    BATCH_SIZE:
+        TRAIN: 5
+        TEST: 1
+    ITER_MAX: 20000
+    ITER_SIZE: 2
+    ITER_SAVE: 5000
+    ITER_TB: 100
+    LR_DECAY: 10
+    LR: 2.5e-4
+    MOMENTUM: 0.9
+    OPTIMIZER: sgd
+    POLY_POWER: 0.9
+    WEIGHT_DECAY: 5.0e-4
+    AVERAGE_LOSS: 20
+
+CRF:
+    ITER_MAX: 10
+    POS_W: 3
+    POS_XY_STD: 1
+    BI_W: 4
+    BI_XY_STD: 67
+    BI_RGB_STD: 3

DeepLab-V2-PyTorch/configs/voc12_2gpu.yaml

@@ -0,0 +1,58 @@
+EXP:
+    ID: voc12
+    OUTPUT_DIR: data
+
+DATASET:
+    NAME: vocaug
+    ROOT: /data/DB/
+    LABELS: ./data/datasets/voc12/labels.txt
+    N_CLASSES: 21
+    IGNORE_LABEL: 255
+    SCALES: [0.5, 0.75, 1.0, 1.25, 1.5]
+    SPLIT:
+        TRAIN: train_cls
+        VAL: val_cls
+        TEST: test
+
+DATALOADER:
+    NUM_WORKERS: 6
+
+IMAGE:
+    MEAN:
+        R: 122.675
+        G: 116.669
+        B: 104.008
+    SIZE:
+        BASE: # None
+        TRAIN: 321
+        TEST: 513
+
+MODEL:
+    NAME: DeepLabV2_ResNet101_MSC
+    N_BLOCKS: [3, 4, 23, 3]
+    ATROUS_RATES: [6, 12, 18, 24]
+    INIT_MODEL: data/models/coco/deeplabv1_resnet101/caffemodel/deeplabv1_resnet101-coco.pth
+
+SOLVER:
+    BATCH_SIZE:
+        TRAIN: 10
+        TEST: 1
+    ITER_MAX: 10000
+    ITER_SIZE: 2
+    ITER_SAVE: 2000
+    ITER_TB: 50
+    LR_DECAY: 10
+    LR: 2.5e-4
+    MOMENTUM: 0.9
+    OPTIMIZER: sgd
+    POLY_POWER: 0.9
+    WEIGHT_DECAY: 5.0e-4
+    AVERAGE_LOSS: 20
+
+CRF:
+    ITER_MAX: 10
+    POS_W: 3
+    POS_XY_STD: 1
+    BI_W: 4
+    BI_XY_STD: 67
+    BI_RGB_STD: 3

DeepLab-V2-PyTorch/configs/voc12_4gpu.yaml

@@ -0,0 +1,58 @@
+EXP:
+    ID: voc12
+    OUTPUT_DIR: data
+
+DATASET:
+    NAME: vocaug
+    ROOT: /data/DB/VOCdevkit
+    LABELS: ./data/datasets/voc12/labels.txt
+    N_CLASSES: 21
+    IGNORE_LABEL: 255
+    SCALES: [0.5, 0.75, 1.0, 1.25, 1.5]
+    SPLIT:
+        TRAIN: train_cls
+        VAL: val_cls
+        TEST: test
+
+DATALOADER:
+    NUM_WORKERS: 12
+
+IMAGE:
+    MEAN:
+        R: 122.675
+        G: 116.669
+        B: 104.008
+    SIZE:
+        BASE: # None
+        TRAIN: 321
+        TEST: 513
+
+MODEL:
+    NAME: DeepLabV2_ResNet101_MSC
+    N_BLOCKS: [3, 4, 23, 3]
+    ATROUS_RATES: [6, 12, 18, 24]
+    INIT_MODEL: data/models/coco/deeplabv1_resnet101/caffemodel/deeplabv1_resnet101-coco.pth
+
+SOLVER:
+    BATCH_SIZE:
+        TRAIN: 20
+        TEST: 1
+    ITER_MAX: 5000
+    ITER_SIZE: 2
+    ITER_SAVE: 1000
+    ITER_TB: 50
+    LR_DECAY: 10
+    LR: 2.5e-4
+    MOMENTUM: 0.9
+    OPTIMIZER: sgd
+    POLY_POWER: 0.9
+    WEIGHT_DECAY: 5.0e-4
+    AVERAGE_LOSS: 20
+
+CRF:
+    ITER_MAX: 10
+    POS_W: 3
+    POS_XY_STD: 1
+    BI_W: 4
+    BI_XY_STD: 67
+    BI_RGB_STD: 3