def test_training():
"""Test only that no error raised."""
config = EasyDict()
config.NETWORK_CLASS = YoloV1
config.DATASET_CLASS = LmThingsOnATable
config.IS_DEBUG = False
config.IMAGE_SIZE = [70, 70]
config.BATCH_SIZE = 4
config.TEST_STEPS = 1
config.MAX_STEPS = 2
config.SAVE_CHECKPOINT_STEPS = 1
config.KEEP_CHECKPOINT_MAX = 5
config.SUMMARISE_STEPS = 1
config.IS_PRETRAIN = False
config.TASK = Tasks.OBJECT_DETECTION
# network model config
config.NETWORK = EasyDict()
config.NETWORK.IMAGE_SIZE = config.IMAGE_SIZE
config.NETWORK.BATCH_SIZE = config.BATCH_SIZE
# daasegt config
config.DATASET = EasyDict()
config.DATASET.PRE_PROCESSOR = ResizeWithGtBoxes(config.IMAGE_SIZE)
config.DATASET.BATCH_SIZE = config.BATCH_SIZE
environment.init("test_yolov_1")
prepare_dirs(recreate=True)
start_training(config)
def test_training():
"""Test only that no error raised."""
config = EasyDict()
config.NETWORK_CLASS = YoloV2
config.DATASET_CLASS = LmThingsOnATable
config.IS_DEBUG = False
config.IMAGE_SIZE = [128, 160]
config.BATCH_SIZE = 2
config.TEST_STEPS = 1
config.MAX_STEPS = 2
config.SAVE_STEPS = 1
config.SUMMARISE_STEPS = 1
config.IS_PRETRAIN = False
config.IS_DISTRIBUTION = False
# network model config
config.NETWORK = EasyDict()
config.NETWORK.OPTIMIZER_CLASS = tf.train.AdamOptimizer
config.NETWORK.OPTIMIZER_KWARGS = {"learning_rate": 0.001}
config.NETWORK.IMAGE_SIZE = config.IMAGE_SIZE
config.NETWORK.BATCH_SIZE = config.BATCH_SIZE
config.NETWORK.DATA_FORMAT = "NCHW"
# daasegt config
config.DATASET = EasyDict()
config.DATASET.PRE_PROCESSOR = ResizeWithGtBoxes(config.IMAGE_SIZE)
config.DATASET.BATCH_SIZE = config.BATCH_SIZE
config.DATASET.DATA_FORMAT = "NCHW"
environment.init("test_yolo_v2")
prepare_dirs(recreate=True)
start_training(config)
def test_training():
"""Test only no error raised."""
config = EasyDict()
config.NETWORK_CLASS = Darknet
config.DATASET_CLASS = Dummy
config.IS_DEBUG = False
config.IMAGE_SIZE = [28, 14]
config.BATCH_SIZE = 2
config.TEST_STEPS = 1
config.MAX_STEPS = 2
config.SAVE_CHECKPOINT_STEPS = 1
config.KEEP_CHECKPOINT_MAX = 5
config.SUMMARISE_STEPS = 1
config.IS_PRETRAIN = False
config.TASK = Tasks.CLASSIFICATION
# network model config
config.NETWORK = EasyDict()
config.NETWORK.OPTIMIZER_CLASS = tf.train.AdamOptimizer
config.NETWORK.OPTIMIZER_KWARGS = {"learning_rate": 0.001}
config.NETWORK.IMAGE_SIZE = config.IMAGE_SIZE
config.NETWORK.BATCH_SIZE = config.BATCH_SIZE
# daasegt config
config.DATASET = EasyDict()
config.DATASET.PRE_PROCESSOR = Resize(config.IMAGE_SIZE)
config.DATASET.BATCH_SIZE = config.BATCH_SIZE
environment.init("test_darknet")
prepare_dirs(recreate=True)
start_training(config)
def test_training():
"""Verify only that no error raised."""
config = EasyDict()
config.NETWORK_CLASS = LMBiSeNet
config.DATASET_CLASS = DummyCamvid
config.IS_DEBUG = False
config.IMAGE_SIZE = [128, 160]
config.BATCH_SIZE = 2
config.TEST_STEPS = 1
config.MAX_STEPS = 2
config.SAVE_STEPS = 1
config.SUMMARISE_STEPS = 1
config.IS_PRETRAIN = False
config.IS_DISTRIBUTION = False
# network model config
config.NETWORK = EasyDict()
config.NETWORK.OPTIMIZER_CLASS = tf.train.AdamOptimizer
config.NETWORK.OPTIMIZER_KWARGS = {"learning_rate": 0.001}
config.NETWORK.IMAGE_SIZE = config.IMAGE_SIZE
config.NETWORK.BATCH_SIZE = config.BATCH_SIZE
config.NETWORK.DATA_FORMAT = "NHWC"
# daasegt config
config.DATASET = EasyDict()
config.DATASET.PRE_PROCESSOR = Resize(config.IMAGE_SIZE)
config.DATASET.BATCH_SIZE = config.BATCH_SIZE
config.DATASET.DATA_FORMAT = "NHWC"
environment.init("test_lm_bisenet")
prepare_dirs(recreate=True)
start_training(config)
def test_training():
"""Test only that no error raised."""
config = EasyDict()
config.NETWORK_CLASS = YoloV2
config.DATASET_CLASS = Pascalvoc2007
config.IS_DEBUG = False
config.IMAGE_SIZE = [128, 160]
config.BATCH_SIZE = 2
config.TEST_STEPS = 1
config.MAX_STEPS = 2
config.SAVE_CHECKPOINT_STEPS = 1
config.KEEP_CHECKPOINT_MAX = 5
config.SUMMARISE_STEPS = 1
config.IS_PRETRAIN = False
config.TASK = Tasks.OBJECT_DETECTION
# network model config
config.NETWORK = EasyDict()
config.NETWORK.OPTIMIZER_CLASS = tf.train.AdamOptimizer
config.NETWORK.OPTIMIZER_KWARGS = {"learning_rate": 0.001}
config.NETWORK.IMAGE_SIZE = config.IMAGE_SIZE
config.NETWORK.BATCH_SIZE = config.BATCH_SIZE
config.NETWORK.DATA_FORMAT = "NHWC"
# dataset config
config.DATASET = EasyDict()
config.DATASET.PRE_PROCESSOR = ResizeWithGtBoxes(config.IMAGE_SIZE)
config.DATASET.BATCH_SIZE = config.BATCH_SIZE
config.DATASET.DATA_FORMAT = "NHWC"
environment.init("test_yolo_v2")
prepare_dirs(recreate=True)
start_training(config)
def test_quantize_training():
"""Test only that no error raised."""
config = EasyDict()
config.NETWORK_CLASS = FlowNetSV1Quantized
config.DATASET_CLASS = FlyingChairs
config.IS_DEBUG = False
config.IMAGE_SIZE = [384, 512]
config.BATCH_SIZE = 8
config.TEST_STEPS = 1
config.MAX_STEPS = 2
config.SAVE_CHECKPOINT_STEPS = 1
config.KEEP_CHECKPOINT_MAX = 5
config.SUMMARISE_STEPS = 1
config.IS_PRETRAIN = False
config.IS_DISTRIBUTION = False
# network model config
config.NETWORK = EasyDict()
config.NETWORK.OPTIMIZER_CLASS = tf.train.AdamOptimizer
config.NETWORK.OPTIMIZER_KWARGS = {"learning_rate": 0.001}
config.NETWORK.IMAGE_SIZE = config.IMAGE_SIZE
config.NETWORK.BATCH_SIZE = config.BATCH_SIZE
config.NETWORK.DATA_FORMAT = "NHWC"
config.NETWORK.ACTIVATION_QUANTIZER = linear_mid_tread_half_quantizer
config.NETWORK.ACTIVATION_QUANTIZER_KWARGS = {'bit': 2, 'max_value': 2.0}
config.NETWORK.WEIGHT_QUANTIZER = binary_channel_wise_mean_scaling_quantizer
config.NETWORK.WEIGHT_QUANTIZER_KWARGS = {}
# dataset config
config.DATASET = EasyDict()
config.DATASET.PRE_PROCESSOR = None
config.DATASET.BATCH_SIZE = config.BATCH_SIZE
config.DATASET.DATA_FORMAT = "NHWC"
config.DATASET.VALIDATION_RATE = 0.2
config.DATASET.VALIDATION_SEED = 2019
config.DATASET.AUGMENTOR = Sequence([
# Geometric transformation
FlipLeftRight(0.5),
FlipTopBottom(0.5),
Translate(-0.2, 0.2),
Rotate(-17, +17),
Scale(1.0, 2.0),
# Pixel-wise augmentation
Brightness(0.8, 1.2),
Contrast(0.2, 1.4),
Color(0.5, 2.0),
Gamma(0.7, 1.5),
# Hue(-128.0, 128.0),
GaussianNoise(0.0, 10.0)
])
config.DATASET.PRE_PROCESSOR = Sequence([
DevideBy255(),
])
environment.init("test_flownet_s_v1_quantize")
prepare_dirs(recreate=True)
start_training(config)
def SqueezeSegV2Config():
mc = EasyDict()
mc.CLASSES = [
'Road', 'Sidewalk', 'Building', 'Pole', 'Vegetation', 'Person',
'TwoWheeler', 'Car', 'Truck', 'Bus', "None"
]
mc.NUM_CLASS = len(mc.CLASSES)
mc.CLS_2_ID = dict(zip(mc.CLASSES, range(len(mc.CLASSES))))
mc.CLS_LOSS_WEIGHT = np.array(
[1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0])
mc.CLS_COLOR_MAP = np.array([
[128, 64, 128], # Road
[244, 35, 232], # Sidewalk
[70, 70, 70], # Building
[153, 153, 153], # Pole
[107, 142, 35], # Vegetation
[220, 20, 60], # Person
[255, 0, 0], # Two Wheeler
[0, 0, 142], # Car
[0, 0, 70], # Truck
[0, 60, 100], # Bus
[0, 0, 0] # None
]) / 255.0
# Input Shape
mc.BATCH_SIZE = 8
mc.AZIMUTH_LEVEL = 240
mc.ZENITH_LEVEL = 32
mc.NUM_FEATURES = 6
# Loss
mc.FOCAL_GAMMA = 2.0
mc.CLS_LOSS_COEF = 15.0
mc.DENOM_EPSILON = 1e-12 # small value used in denominator to prevent division by 0
# Gradient Decent
mc.LEARNING_RATE = 0.05
mc.LR_DECAY_STEPS = 500
mc.LR_DECAY_FACTOR = 0.9
mc.MAX_GRAD_NORM = 100.0
# Network
mc.L2_WEIGHT_DECAY = 0.05
mc.DROP_RATE = 0.1
mc.BN_MOMENTUM = 0.9
mc.REDUCTION = 16
# Dataset
mc.DATA_AUGMENTATION = True
mc.RANDOM_FLIPPING = True
mc.RANDOM_SHIFT = True
# x, y, z, intensity, distance
mc.INPUT_MEAN = np.array([[[24.810, 0.819, 0.000, 16.303, 25.436]]])
mc.INPUT_STD = np.array([[[30.335, 7.807, 2.058, 25.208, 30.897]]])
return mc
def test_training():
"""Test only no error raised."""
config = EasyDict()
config.NETWORK_CLASS = LmSinglePoseV1Quantize
config.DATASET_CLASS = MscocoSinglePersonKeypoints
config.IS_DEBUG = False
config.IMAGE_SIZE = [160, 160]
config.BATCH_SIZE = 2
config.TEST_STEPS = 1
config.MAX_STEPS = 2
config.SAVE_CHECKPOINT_STEPS = 1
config.KEEP_CHECKPOINT_MAX = 5
config.SUMMARISE_STEPS = 1
config.IS_PRETRAIN = False
config.IS_DISTRIBUTION = False
config.TASK = Tasks.KEYPOINT_DETECTION
# network model config
config.NETWORK = EasyDict()
config.NETWORK.OPTIMIZER_CLASS = tf.train.AdamOptimizer
config.NETWORK.OPTIMIZER_KWARGS = {"learning_rate": 0.001}
config.NETWORK.IMAGE_SIZE = config.IMAGE_SIZE
config.NETWORK.BATCH_SIZE = config.BATCH_SIZE
config.NETWORK.ACTIVATION_QUANTIZER = linear_mid_tread_half_quantizer
config.NETWORK.ACTIVATION_QUANTIZER_KWARGS = {
'bit': 2,
'max_value': 2.0
}
config.NETWORK.WEIGHT_QUANTIZER = binary_channel_wise_mean_scaling_quantizer
config.NETWORK.WEIGHT_QUANTIZER_KWARGS = {}
# daasegt config
config.DATASET = EasyDict()
config.DATASET.PRE_PROCESSOR = Sequence([
ResizeWithJoints(image_size=config.IMAGE_SIZE),
JointsToGaussianHeatmap(image_size=config.IMAGE_SIZE, stride=2),
DivideBy255()])
config.DATASET.BATCH_SIZE = config.BATCH_SIZE
environment.init("test_lm_single_pose_v1")
prepare_dirs(recreate=True)
start_training(config, profile_step=1)
def test_training():
"""Test only no error raised."""
config = EasyDict()
config.NETWORK_CLASS = YoloV2Quantize
config.DATASET_CLASS = Pascalvoc2007
config.IS_DEBUG = False
config.IMAGE_SIZE = [128, 160]
config.BATCH_SIZE = 2
config.TEST_STEPS = 1
config.MAX_STEPS = 2
config.SAVE_CHECKPOINT_STEPS = 1
config.KEEP_CHECKPOINT_MAX = 5
config.SUMMARISE_STEPS = 1
config.IS_PRETRAIN = False
config.TASK = Tasks.OBJECT_DETECTION
# network model config
config.NETWORK = EasyDict()
config.NETWORK.OPTIMIZER_CLASS = tf.train.AdamOptimizer
config.NETWORK.OPTIMIZER_KWARGS = {"learning_rate": 0.001}
config.NETWORK.IMAGE_SIZE = config.IMAGE_SIZE
config.NETWORK.BATCH_SIZE = config.BATCH_SIZE
config.NETWORK.ACTIVATION_QUANTIZER = linear_mid_tread_half_quantizer
config.NETWORK.ACTIVATION_QUANTIZER_KWARGS = {
'bit': 2,
'max_value': 2.0
}
config.NETWORK.WEIGHT_QUANTIZER = binary_channel_wise_mean_scaling_quantizer
config.NETWORK.WEIGHT_QUANTIZER_KWARGS = {}
# daasegt config
config.DATASET = EasyDict()
config.DATASET.PRE_PROCESSOR = ResizeWithGtBoxes(config.IMAGE_SIZE)
config.DATASET.BATCH_SIZE = config.BATCH_SIZE
environment.init("test_yolov_2_quantize")
prepare_dirs(recreate=True)
start_training(config)
def test_training():
"""Test only that no error raised."""
config = EasyDict()
config.NETWORK_CLASS = LmResnetQuantize
config.DATASET_CLASS = Dummy
config.IS_DEBUG = False
config.IMAGE_SIZE = [32, 32]
config.BATCH_SIZE = 2
config.TEST_STEPS = 1
config.MAX_STEPS = 2
config.SAVE_CHECKPOINT_STEPS = 1
config.KEEP_CHECKPOINT_MAX = 5
config.SUMMARISE_STEPS = 1
config.IS_PRETRAIN = False
config.TASK = Tasks.CLASSIFICATION
# network model config
config.NETWORK = EasyDict()
config.NETWORK.OPTIMIZER_CLASS = tf.train.AdamOptimizer
config.NETWORK.OPTIMIZER_KWARGS = {"learning_rate": 0.001}
config.NETWORK.IMAGE_SIZE = config.IMAGE_SIZE
config.NETWORK.BATCH_SIZE = config.BATCH_SIZE
config.NETWORK.ACTIVATION_QUANTIZER = linear_mid_tread_half_quantizer
config.NETWORK.ACTIVATION_QUANTIZER_KWARGS = {
'bit': 2,
'max_value': 2
}
config.NETWORK.WEIGHT_QUANTIZER = binary_mean_scaling_quantizer
config.NETWORK.WEIGHT_QUANTIZER_KWARGS = {}
# dataset config
config.DATASET = EasyDict()
config.DATASET.PRE_PROCESSOR = Resize(config.IMAGE_SIZE)
config.DATASET.BATCH_SIZE = config.BATCH_SIZE
environment.init("test_lm_resnet_quantize")
prepare_dirs(recreate=True)
start_training(config)
# 'mnist': MNIST
# 'cifar10' CIFAR-10
__C.DATABASE_NAME = 'cifar10'
# Training version
# Set None to auto generate version
__C.VERSION = None
# Learning rate
__C.LEARNING_RATE = 0.001
# Epochs
__C.EPOCHS = 50
# Batch size
__C.BATCH_SIZE = 256
# Setting test set as validation when preprocessing data
__C.DPP_TEST_AS_VALID = False
# ===========================================
# # Model Architecture #
# ===========================================
# -------------------------------------------
# Classification
# Parameters of margin loss
# default: {'m_plus': 0.9, 'm_minus': 0.1, 'lambda_': 0.5}
__C.MARGIN_LOSS_PARAMS = {'m_plus': 0.9, 'm_minus': 0.1, 'lambda_': 0.5}
# SUMMARISE_STEPS = 2
# IS_DEBUG = True
PRE_PROCESSOR = Sequence([Resize(size=IMAGE_SIZE), PerImageStandardization()])
POST_PROCESSOR = None
NETWORK = EasyDict()
NETWORK.OPTIMIZER_CLASS = tf.train.MomentumOptimizer
NETWORK.OPTIMIZER_KWARGS = {"momentum": 0.9}
NETWORK.LEARNING_RATE_FUNC = tf.train.piecewise_constant
NETWORK.LEARNING_RATE_KWARGS = {
"values": [0.1, 0.01, 0.001, 0.0001],
"boundaries": [40000, 60000, 80000],
}
NETWORK.IMAGE_SIZE = IMAGE_SIZE
NETWORK.BATCH_SIZE = BATCH_SIZE
NETWORK.DATA_FORMAT = DATA_FORMAT
NETWORK.WEIGHT_DECAY_RATE = 0.0001
NETWORK.ACTIVATION_QUANTIZER = linear_mid_tread_half_quantizer
NETWORK.ACTIVATION_QUANTIZER_KWARGS = {'bit': 2, 'max_value': 2}
NETWORK.WEIGHT_QUANTIZER = binary_mean_scaling_quantizer
NETWORK.WEIGHT_QUANTIZER_KWARGS = {}
# dataset
DATASET = EasyDict()
DATASET.BATCH_SIZE = BATCH_SIZE
DATASET.DATA_FORMAT = DATA_FORMAT
DATASET.PRE_PROCESSOR = PRE_PROCESSOR
DATASET.AUGMENTOR = Sequence([
Pad(2),
Crop(size=IMAGE_SIZE),
# Learning rate __C.LEARNING_RATE = 0.001 # Learning rate with exponential decay # Use learning rate decay __C.LR_DECAY = True # Decay steps __C.LR_DECAY_STEPS = 2000 # Exponential decay rate __C.LR_DECAY_RATE = 0.96 # Epochs __C.EPOCHS = 20 # Batch size __C.BATCH_SIZE = 512 # Data format # 'NCHW': (batch, channel, height, width) # 'NHWC': (batch, height, width, channel) __C.DATA_FORMAT = 'NCHW' # =========================================== # # Preprocessing # # =========================================== # Setting test set as validation when preprocessing data __C.DPP_TEST_AS_VALID = True # Rate of train-test split __C.TEST_SIZE = 0.2
# coding:utf-8 from easydict import EasyDict config = EasyDict() config.BATCH_SIZE = 384 config.CLS_OHEM = True config.CLS_OHEM_RATIO = 0.7 config.BBOX_OHEM = False config.BBOX_OHEM_RATIO = 0.7 config.GPU = '2,3' config.VISIBLE_GPU = '2,3' config.EPS = 1e-14 config.LR_EPOCH = [6, 14, 20] # [6, 14, 20]
YOLO_CONFIG.USE_FOCAL_LOSS = False YOLO_CONFIG.USE_LABEL_SMOOTH = False YOLO_CONFIG.TRAIN_PATH = '/home/huydao/Source/LogoDetection/data/metadata/train.txt' YOLO_CONFIG.VAL_PATH = '/home/huydao/Source/LogoDetection/data/metadata/val.txt' YOLO_CONFIG.N_CLASSES = 1 YOLO_CONFIG.NMS_TOPK = 150 YOLO_CONFIG.NMS_THRESHOLD = 0.45 YOLO_CONFIG.SCORE_THRESHOLD = 0.01 YOLO_CONFIG.RESTORE_PATH = '/home/huydao/Source/LogoDetection/data/weights/yolov3.ckpt' YOLO_CONFIG.SAVE_DIR = '/home/huydao/Source/LogoDetection/checkpoints' YOLO_CONFIG.BATCH_SIZE = 8 YOLO_CONFIG.TOTAL_EPOCHS = 10 YOLO_CONFIG.TRAIN_EVALUATION_STEP = 10 YOLO_CONFIG.VAL_EVALUATION_STEP = 2 YOLO_CONFIG.SAVE_EPOCH = 10 YOLO_CONFIG.BATCH_NORM_DECAY = 0.99 WEIGHT_DECAY = 5e-4 YOLO_CONFIG.CLASSES_PATH = '' YOLO_CONFIG.ANCHORS_PATH = '' YOLO_CONFIG.TRAIN_EXAMPLES = len(open(YOLO_CONFIG.TRAIN_PATH,'r').readlines()) YOLO_CONFIG.VAL_EXAMPLES = len(open(YOLO_CONFIG.VAL_PATH,'r').readlines()) YOLO_CONFIG.TRAIN_BATCH_NUM = int(math.ceil(float(YOLO_CONFIG.TRAIN_EXAMPLES)/YOLO_CONFIG.BATCH_SIZE))
for item in NAME_TO_ID.items():
ID_TO_NAME[item[1]] = item[0]
"""
Logging settings
"""
cfg.LOG_DIR = os.path.join(cfg.ROOT_DIR, 'runs', cfg.RUN_NAME)
cfg.CLEAR_HISTORY = conf['clear_history']
cfg.PRINT_EVERY = conf['print_every']
"""
Training settings
"""
cfg.NUM_EPOCHS = conf['num_epochs']
cfg.BATCH_SIZE = conf['batch_size']
cfg.DEVICE = torch.device('cuda:1' if torch.cuda.is_available() else 'cpu')
cfg.PATIENCE = conf['patience']
cfg.RESUME = conf['resume']
cfg.NUM_WORKERS = conf['num_workers']
cfg.LEARNING_RATE = conf['learning_rate']
cfg.WEIGHT_DECAY = conf['weight_decay']
"""
Model saving
"""
cfg.CHECKPOINT_DIR = os.path.join(cfg.ROOT_DIR, 'checkpoint')
cfg.BEST_MODEL_PATH = os.path.join(cfg.CHECKPOINT_DIR, 'best_model.pth')
cfg.BEST_ACC_PATH = os.path.join(cfg.CHECKPOINT_DIR, 'best_acc.txt')
cfg.CHECKPOINT_PATH = os.path.join(cfg.CHECKPOINT_DIR, cfg.RUN_NAME + '.tar')
"""
Clear history
conf.LR_BG = 1e-3 conf.EPOCH_SIZE_BG = 8000 conf.BATCH_SIZE_BG = 15 conf.LD_DECAY_BG = 2000 conf.BG_CUE_FILE = "bg_cue_initsec.p" conf.SALIENCY_TH_BG = 0.1 #training params for SEC models conf.CUE_FILE_INITSEC = "sec_cue.p" conf.INPUT_SIZE_SEC = 320 conf.DOWN_SAMPLE_SEC = 8 #network resolution conf.Q_FG = 0.996 conf.Q_BG = 0.999 conf.LR = 1e-3 conf.LR_DECAY = 2000 conf.MAX_EPOCH = 8 conf.EPOCH_SIZE = 1000 conf.BATCH_SIZE = 15 conf.WD = 5e-4 conf.MOMENTUM = 0.9 conf.WORKSPACE = 512 conf.CPU_WORKER_NUM = 8 conf.CRF_POS_XY_STD = 3 conf.CRF_POS_W = 3 conf.CRF_BI_RGB_STD = 10 conf.CRF_BI_XY_STD = 80 conf.CRF_BI_W = 10
from utils import ShapeSpec
cfg = ED()
# TODO: Most hyperparameters be hard code(so I can test it easy), it all will be added to this file later(maybe).
# trainer
# epoch, batch_size, so on.
# base info
cfg.NUM_CLASSES = 1
cfg.EPOCH = 100
cfg.lr = 1e-4
# data
cfg.ROOT = 'datasets/wgisd'
cfg.BATCH_SIZE = 3
cfg.NUM_WORKERS = 8
cfg.RESIZE = (800, 1280)
# backbone
cfg.BACKBONE_DEPTH = 101
cfg.BACKBONE_OUTPUT_SHAPE = {
'res2': ShapeSpec(channels=256, height=None, width=None, stride=4),
'res3': ShapeSpec(channels=512, height=None, width=None, stride=8),
'res4': ShapeSpec(channels=1024, height=None, width=None, stride=16),
'res5': ShapeSpec(channels=2048, height=None, width=None, stride=32)
}
# MASK
cfg.CLS_AGNOSTIC_MASK = False
# Learning rate __C.LEARNING_RATE = 0.001 # Learning rate with exponential decay # Use learning rate decay __C.LR_DECAY = True # Decay steps __C.LR_DECAY_STEPS = 2000 # Exponential decay rate __C.LR_DECAY_RATE = 0.96 # Epochs __C.EPOCHS = 50 # Batch size __C.BATCH_SIZE = 2048 # =========================================== # # Preprocessing # # =========================================== # Setting test set as validation when preprocessing data __C.DPP_TEST_AS_VALID = True # Rate of train-test split __C.TEST_SIZE = 0.2 # Rate of train-validation split __C.VALID_SIZE = 0.1 # Resize images
