INPUT_SIZE = 416
BATCH_SIZE = 1
EPOCHS = 20
LR = 0.001
SHUFFLE_SIZE = 1
weights_path = "/home/yang/test/yolov3.weights"
sess = tf.Session()
classes = utils.read_coco_names('./data/coco.names')
num_classes = len(classes)
file_pattern = "./data/train_data/quick_train_data/tfrecords/quick_train_data*.tfrecords"
anchors = utils.get_anchors('./data/yolo_anchors.txt')
dataset = tf.data.TFRecordDataset(filenames=tf.gfile.Glob(file_pattern))
dataset = dataset.map(utils.parser(anchors, num_classes).parser_example,
num_parallel_calls=10)
dataset = dataset.repeat().shuffle(SHUFFLE_SIZE).batch(BATCH_SIZE).prefetch(
BATCH_SIZE)
iterator = dataset.make_one_shot_iterator()
example = iterator.get_next()
images, *y_true = example
model = yolov3.yolov3(num_classes)
with tf.variable_scope('yolov3'):
y_pred = model.forward(images, is_training=False)
loss = model.compute_loss(y_pred, y_true)
y_pred = model.predict(y_pred)
load_ops = utils.load_weights(tf.global_variables(scope='yolov3'),
weights_path)
sess.run(load_ops)
INPUT_SIZE = 416
BATCH_SIZE = 16
EPOCHS = 5000000
LR = 0.0001
SHUFFLE_SIZE = 10000
sess = tf.Session()
classes = utils.read_coco_names('./data/coco.names')
num_classes = len(classes)
# file_pattern = "../COCO/tfrecords/coco*.tfrecords"
file_pattern = "./data/train_data/quick_train_data/tfrecords/quick_train_data*.tfrecords"
anchors = utils.get_anchors('./data/yolo_anchors.txt')
is_training = tf.placeholder(dtype=tf.bool, name="phase_train")
dataset = tf.data.TFRecordDataset(filenames = tf.gfile.Glob(file_pattern))
dataset = dataset.map(utils.parser(anchors, num_classes).parser_example, num_parallel_calls = 10)
dataset = dataset.repeat().shuffle(SHUFFLE_SIZE).batch(BATCH_SIZE).prefetch(BATCH_SIZE)
iterator = dataset.make_one_shot_iterator()
example = iterator.get_next()
images, *y_true = example
model = yolov3.yolov3(num_classes)
with tf.variable_scope('yolov3'):
y_pred = model.forward(images, is_training=is_training)
loss = model.compute_loss(y_pred, y_true)
y_pred = model.predict(y_pred)
optimizer = tf.train.MomentumOptimizer(LR, momentum=0.9)
train_op = optimizer.minimize(loss[0])
saver = tf.train.Saver(max_to_keep=2)