def parse_train_config(config=None):
config = {} if not config else config
c = AttrDict()
c.DATASET_ROOT = config.get("DATASET_ROOT", DATASET_ROOT)
c.JSON_PATH = config.get("JSON_PATH", "train.json")
c.BATCH_SIZE = config.get("BATCH_SIZE", BATCH_SIZE)
c.IMAGE_SIZE = config.get("IMAGE_SIZE", IMAGE_SIZE)
c.WORKERS = config.get("WORKERS", WORKERS)
c.PIN_MEMORY = config.get("PIN_MEMORY", PIN_MEMORY)
c.SHUFFLE = config.get("SHUFFLE", True)
c.LEARNING_RATE = config.get("LEARNING_RATE", LEARNING_RATE)
c.MOMENTUM = config.get("MOMENTUM", MOMENTUM)
c.DAMPENING = config.get("DAMPENING", DAMPENING)
c.BETAS = config.get("BETAS", BETAS)
c.EPS = config.get("EPS", EPS)
c.WEIGHT_DECAY = config.get("WEIGHT_DECAY", WEIGHT_DECAY)
c.MILESTONES = config.get("MILESTONES", MILESTONES)
c.GAMMA = config.get("GAMMA", GAMMA)
c.NUM_EPOCHS = config.get("NUM_EPOCHS", NUM_EPOCHS)
c.TEST = config.get("TEST", TEST)
c.OUT_PATH = config.get("OUT_PATH", OUT_PATH)
c.LOAD_MODEL = config.get("LOAD_MODEL", LOAD_MODEL)
c.SAVE_MODEL = config.get("SAVE_MODEL", SAVE_MODEL)
c.CHECKPOINT_FILE = config.get("CHECKPOINT_FILE", CHECKPOINT_FILE)
return c
def parse_detect_config(config=None):
config = {} if not config else config
c = AttrDict()
c.JSON = config.get("JSON", JSON)
c.IMAGE_SIZE = config.get("IMAGE_SIZE", IMAGE_SIZE)
c.CHECKPOINT_FILE = config.get("CHECKPOINT_FILE", CHECKPOINT_FILE)
return c
def parse_test_config(config=None):
config = {} if not config else config
c = AttrDict()
c.DATASET_ROOT = config.get("DATASET_ROOT", DATASET_ROOT)
c.JSON_PATH = config.get("JSON_PATH", "test.json")
c.BATCH_SIZE = config.get("BATCH_SIZE", BATCH_SIZE)
c.IMAGE_SIZE = config.get("IMAGE_SIZE", IMAGE_SIZE)
c.WORKERS = config.get("WORKERS", WORKERS)
c.PIN_MEMORY = config.get("PIN_MEMORY", PIN_MEMORY)
c.SHUFFLE = config.get("SHUFFLE", False)
c.OUT_PATH = config.get("OUT_PATH", OUT_PATH)
c.LOAD_MODEL = config.get("LOAD_MODEL", True)
c.CHECKPOINT_FILE = config.get("CHECKPOINT_FILE", CHECKPOINT_FILE)
return c
def get_params():
checkpoint_dir = '/Users/Nolsigan/PycharmProjects/rlntm-tensorflow/checkpoints'
max_length = 6
rnn_cell = rnn.BasicLSTMCell
rnn_hidden = 128
learning_rate = 0.003
optimizer = tf.train.AdamOptimizer()
gradient_clipping = 5
batch_size = 100
epochs = 30
epoch_size = 100
num_symbols = 10
dup_factor = 2
mem_dim = 128
mem_move_table = [-1, 0, 1]
in_move_table = [-1, 0, 1]
out_move_table = [0, 1]
return AttrDict(**locals())
# Config ############################################################################## from __future__ import absolute_import from __future__ import division from __future__ import print_function from __future__ import unicode_literals import os import re import torch from attr_dict import AttrDict __C = AttrDict() cfg = __C __C.GLOBAL_RANK = 0 __C.EPOCH = 0 # Absolute path to a location to keep some large files, not in this dir. __C.ASSETS_PATH = '/home/dcg-adlr-atao-data.cosmos277/assets' # Use class weighted loss per batch to increase loss for low pixel count classes per batch __C.BATCH_WEIGHTING = False # Border Relaxation Count __C.BORDER_WINDOW = 1 # Number of epoch to use before turn off border restriction __C.REDUCE_BORDER_EPOCH = -1 # Comma Seperated List of class id to relax __C.STRICTBORDERCLASS = None
