BN_INIT_DECAY = 0.5
BN_DECAY_DECAY_RATE = 0.5
BN_DECAY_DECAY_STEP = float(DECAY_STEP)
BN_DECAY_CLIP = 0.99
HOSTNAME = socket.gethostname()
NUM_CLASSES = 50
# Shapenet official train/test split
DATA_PATH = os.path.join(
ROOT_DIR, 'data', 'shapenetcore_partanno_segmentation_benchmark_v0_normal')
TRAIN_DATASET = part_dataset_all_normal.PartNormalDataset(
root=DATA_PATH,
npoints=NUM_POINT,
classification=False,
split='trainval',
return_cls_label=True)
TEST_DATASET = part_dataset_all_normal.PartNormalDataset(root=DATA_PATH,
npoints=NUM_POINT,
classification=False,
split='test',
return_cls_label=True)
def log_string(out_str):
LOG_FOUT.write(out_str + '\n')
LOG_FOUT.flush()
print(out_str)
if TEST_ROTATED:
DATA_PATH = os.path.join(DATA_DIR, 'shapenetcore_normal_rotated')
TEST_DATASET = part_dataset_all_normal_rotated.PartNormalDataset(
root=DATA_PATH,
npoints=NUM_POINT,
classification=False,
split='test_rand',
return_cls_label=True,
random_sampling=True)
else:
# Shapenet official train/test split
DATA_PATH = os.path.join(DATA_DIR, 'shapenetcore_v0_normal')
TEST_DATASET = part_dataset_all_normal.PartNormalDataset(
root=DATA_PATH,
npoints=NUM_POINT,
classification=False,
split='test',
return_cls_label=True,
random_sampling=False)
def log_string(out_str):
LOG_FOUT.write(out_str + '\n')
LOG_FOUT.flush()
print(out_str)
def evaluate():
with tf.Graph().as_default():
with tf.device('/gpu:' + str(GPU_INDEX)):
pointclouds_pl, labels_pl, cls_labels_pl = MODEL.eval_placeholder_inputs(
with open(out_file, 'w') as f:
seg = seg[0]
l = len(seg)
for i in range(l):
if seg[i] == 1:
color = [0, 0, 1]
elif seg[i] == 0:
color = [1, 0, 0]
else:
color = [0, 0, 0]
f.write('v %f %f %f %f %f %f\n' % (data[i][0], data[i][1], data[i][2], color[0], color[1], color[2]))
# Shapenet official train/test split
DATA_PATH = os.path.join(Data_Path, 'data', 'shapenetcore_partanno_segmentation_benchmark_v0_normal')
TEST_DATASET = part_dataset_all_normal.PartNormalDataset(root=DATA_PATH, npoints=NUM_POINT, classification=False, split='test')
def evaluate():
with tf.Graph().as_default():
with tf.device('/gpu:' + str(GPU_INDEX)):
pointclouds_pl, labels_pl, cls_labels_pl = MODEL.placeholder_inputs(BATCH_SIZE, NUM_POINT)
is_training_pl = tf.placeholder(tf.bool, shape=())
print(is_training_pl)
print("--- Get model and loss")
pred, _ = MODEL.get_model(pointclouds_pl, cls_labels_pl, is_training_pl)
loss = MODEL.get_loss(pred, labels_pl)
saver = tf.train.Saver()
# Create a session
config = tf.ConfigProto()
os.system('cp train.py %s' % (LOG_DIR)) # bkp of train procedure
LOG_FOUT = open(os.path.join(LOG_DIR, 'log_train.txt'), 'w')
LOG_FOUT.write(str(FLAGS)+'\n')
BN_INIT_DECAY = 0.5
BN_DECAY_DECAY_RATE = 0.5
BN_DECAY_DECAY_STEP = float(DECAY_STEP)
BN_DECAY_CLIP = 0.99
HOSTNAME = socket.gethostname()
NUM_CLASSES = 50
# Shapenet official train/test split
DATA_PATH = os.path.join(DATA_DIR, 'shapenetcore_partanno_segmentation_benchmark_v0_normal')
TRAIN_DATASET = part_dataset_all_normal.PartNormalDataset(root=DATA_PATH, input_cat=INPUT_CAT, npoints=NUM_POINT, classification=False, split='trainval')
TEST_DATASET = part_dataset_all_normal.PartNormalDataset(root=DATA_PATH, input_cat=INPUT_CAT, npoints=NUM_POINT, classification=False, split='test')
def log_string(out_str):
LOG_FOUT.write(out_str+'\n')
LOG_FOUT.flush()
print(out_str)
def get_learning_rate(batch):
learning_rate = tf.train.exponential_decay(
BASE_LEARNING_RATE, # Base learning rate.
batch * BATCH_SIZE, # Current index into the dataset.
DECAY_STEP, # Decay step.
DECAY_RATE, # Decay rate.
staircase=True)
learning_rate = tf.maximum(learning_rate, 0.00001) # CLIP THE LEARNING RATE!