def main(argv=None):
if len(argv) > 1:
model_str = argv[1]
if model_str == 'vgg16':
model = helper.import_module(argv[1], FLAGS.vgg16_model_path)
model_path = FLAGS.vgg16_model_path
elif model == 'vgg19':
model_str = helper.import_module(argv[1], FLAGS.vgg19_model_path)
model_path = FLAGS.vgg19_model_path
else:
model = helper.import_module('model', FLAGS.vgg16_model_path)
model_path = FLAGS.vgg16_model_path
if tf.gfile.Exists(FLAGS.train_dir):
raise ValueError('Train dir exists: ' + FLAGS.train_dir)
tf.gfile.MakeDirs(FLAGS.train_dir)
stats_dir = os.path.join(FLAGS.train_dir, 'stats')
tf.gfile.MakeDirs(stats_dir)
tf.gfile.MakeDirs(os.path.join(FLAGS.train_dir, 'visualize'))
f = open(os.path.join(stats_dir, 'log.txt'), 'w')
sys.stdout = train_helper.Logger(sys.stdout, f)
copyfile(model_path, os.path.join(FLAGS.train_dir, model_str + '.py'))
copyfile(FLAGS.config_path, os.path.join(FLAGS.train_dir, 'config.py'))
print('Experiment dir: ' + FLAGS.train_dir)
train(model)
def main(argv=None):
#ipdb.set_trace()
# import model
model = helper.import_module('model', 'models/model_s.py')
# load image
'''
filename_queue = tf.train.string_input_producer(['rgb_img.png'])
reader = tf.WholeFileReader()
key, value = reader.read(filename_queue)
image = tf.image.decode_png(value)
print (tf.shape(image))
exit()
image = Image.open('rgb_img.png')
image = np.asarray(image)
image = image.astype('uint8')
'''
filename_queue = tf.train.string_input_producer(['rgb_img.png'])
reader = tf.TFRecordReader()
_, serialized_example = reader.read(filename_queue)
features = tf.parse_single_example(serialized_example,
features={
'height':
tf.FixedLenFeature([], tf.int64),
'width':
tf.FixedLenFeature([], tf.int64),
'depth':
tf.FixedLenFeature([], tf.int64),
'num_labels':
tf.FixedLenFeature([], tf.int64),
'img_name':
tf.FixedLenFeature([], tf.string),
'rgb':
tf.FixedLenFeature([], tf.string),
'label_weights':
tf.FixedLenFeature([], tf.string),
'labels':
tf.FixedLenFeature([], tf.string),
})
image = tf.decode_raw(features['rgb'], tf.uint8)
image = tf.reshape(image, shape=[1, 256, 512, 3])
image = tf.to_float(image)
# define path where model checkpoint is stored
resume_path = '/home/stjepan/FAKS/zavrad/VGG_TX1/model.ckpt'
# evaluate one image
evalone(model, resume_path, image)
def main(argv=None):
model = helper.import_module('model', FLAGS.model_path)
if tf.gfile.Exists(FLAGS.train_dir):
raise ValueError('Train dir exists: ' + FLAGS.train_dir)
tf.gfile.MakeDirs(FLAGS.train_dir)
stats_dir = os.path.join(FLAGS.train_dir, 'stats')
tf.gfile.MakeDirs(stats_dir)
tf.gfile.MakeDirs(os.path.join(FLAGS.train_dir, 'visualize'))
f = open(os.path.join(stats_dir, 'log.txt'), 'w')
sys.stdout = train_helper.Logger(sys.stdout, f)
copyfile(FLAGS.model_path, os.path.join(FLAGS.train_dir, 'model.py'))
copyfile(FLAGS.config_path, os.path.join(FLAGS.train_dir, 'config.py'))
print('Experiment dir: ' + FLAGS.train_dir)
train(model)
import eval_helper
import numpy as np
import helper
import sys
from shutil import copyfile
# ffs change this relative path
tf.app.flags.DEFINE_string('config_path', "config/cityscapes.py",
"""Path to experiment config.""")
FLAGS = tf.app.flags.FLAGS
helper.import_module('config', FLAGS.config_path)
class_names = [
'road', 'sidewalk', 'building', 'wall', 'fence', 'pole', 'traffic light',
'traffic sign', 'vegetation', 'terrain', 'sky', 'person', 'rider', 'car',
'truck', 'bus', 'train', 'motorcycle', 'bicycle'
]
def collect_confusion(logits, labels, conf_mat):
predicted_labels = logits.argmax(3).astype(np.int32, copy=False)
num_examples = FLAGS.batch_size * FLAGS.img_height * FLAGS.img_width
predicted_labels = np.resize(predicted_labels, [
num_examples,
])
def main(argv=None):
model = helper.import_module('model', 'models/model_s.py')
resume_path = 'trained_model/model.ckpt'
evaluate(model, resume_path)
def main(argv=None): # pylint: disable=unused-argument
model = helper.import_module('model', os.path.join(FLAGS.model_dir, 'model.py'))
if not tf.gfile.Exists(FLAGS.model_dir):
from datasets.voc2012.dataset import Dataset
np.set_printoptions(linewidth=250)
#DATA_DIR = '/home/kivan/datasets/voc2012_aug/data/'
#split = 'val'
DATA_DIR = '/home/kivan/datasets/VOC2012/test_data'
tf.app.flags.DEFINE_string('model_dir',
#'/home/kivan/datasets/results/tmp/voc2012/25_7_16-38-50/', '')
#'/home/kivan/datasets/results/tmp/voc2012/25_5_22-30-16', '')
FLAGS = tf.app.flags.FLAGS
helper.import_module('config', os.path.join(FLAGS.model_dir, 'config.py'))
def forward_pass(model, save_dir):
img_dir = join(DATA_DIR, 'JPEGImages')
file_path = join(DATA_DIR, 'ImageSets', 'Segmentation', 'test.txt')
fp = open(file_path)
file_list = [line.strip() for line in fp]
save_dir_rgb = join(save_dir, 'rgb')
tf.gfile.MakeDirs(save_dir_rgb)
save_dir_submit = join(save_dir, 'submit')
tf.gfile.MakeDirs(save_dir_submit)
#sess = tf.Session(config=tf.ConfigProto(log_device_placement=FLAGS.log_device_placement))
config = tf.ConfigProto(log_device_placement=FLAGS.log_device_placement)
#config.gpu_options.per_process_gpu_memory_fraction = 0.5 # don't hog all vRAM