def test_pix2pix_fid(model, opt):
opt.phase = 'val'
opt.num_threads = 0
opt.batch_size = 1
opt.serial_batches = True
opt.no_flip = True
opt.load_size = 256
opt.display_id = -1
dataset = create_dataset(opt)
model.model_eval()
result_dir = os.path.join(opt.checkpoints_dir, opt.name, 'test_results')
util.mkdirs(result_dir)
fake_B = {}
for i, data in enumerate(dataset):
model.set_input(data)
with torch.no_grad():
model.forward()
visuals = model.get_current_visuals()
fake_B[data['A_paths'][0]] = visuals['fake_B']
util.save_images(visuals, model.image_paths, result_dir, direction=opt.direction,
aspect_ratio=opt.aspect_ratio)
block_idx = InceptionV3.BLOCK_INDEX_BY_DIM[2048]
inception_model = InceptionV3([block_idx])
inception_model.to(model.device)
inception_model.eval()
npz = np.load(os.path.join(opt.dataroot, 'real_stat_B.npz'))
fid = get_fid(list(fake_B.values()), inception_model, npz, model.device, opt.batch_size)
return fid
def __init__(self, opt, save_dir, filename='loss_log.txt'):
self.display_id = opt.display_id
self.use_html = not opt.no_html
self.win_size = opt.display_winsize
self.save_dir = save_dir
self.name = os.path.basename(self.save_dir)
self.saved = False
self.display_single_pane_ncols = opt.display_single_pane_ncols
# Error plots
self.error_plots = dict()
self.error_wins = dict()
if self.display_id > 0:
import visdom
self.vis = visdom.Visdom(port=opt.display_port)
if self.use_html:
self.web_dir = os.path.join(self.save_dir, 'web')
self.img_dir = os.path.join(self.web_dir, 'images')
print('create web directory %s...' % self.web_dir)
util.mkdirs([self.web_dir, self.img_dir])
self.log_name = os.path.join(self.save_dir, filename)
with open(self.log_name, "a") as log_file:
now = time.strftime("%c")
log_file.write('================ Training Loss (%s) ================\n' % now)
def parse(self, save=True):
if not self.initialized:
self.initialize()
self.opt = self.parser.parse_args()
self.opt.isTrain = self.isTrain # train or test
str_ids = self.opt.gpu_ids.split(',')
self.opt.gpu_ids = []
for str_id in str_ids:
id = int(str_id)
if id >= 0:
self.opt.gpu_ids.append(id)
# # set gpu ids
# if len(self.opt.gpu_ids) > 0:
# torch.cuda.set_device(self.opt.gpu_ids[0])
args = vars(self.opt)
print('------------ Options -------------')
for k, v in sorted(args.items()):
print('%s: %s' % (str(k), str(v)))
print('-------------- End ----------------')
# save to the disk
expr_dir = os.path.join(self.opt.checkpoints_dir, self.opt.name)
util.mkdirs(expr_dir)
if save and not self.opt.continue_train:
file_name = os.path.join(expr_dir, 'opt.txt')
with open(file_name, 'wt') as opt_file:
opt_file.write('------------ Options -------------\n')
for k, v in sorted(args.items()):
opt_file.write('%s: %s\n' % (str(k), str(v)))
opt_file.write('-------------- End ----------------\n')
return self.opt
def parse(self):
if not self.initialized:
self.initialize()
self.opt = self.parser.parse_args()
self.opt.isTrain = self.isTrain # train or test
self.opt.name = self.name # experiment name
#str_ids = self.opt.gpu_ids.split(',')
#self.opt.gpu_ids = []
#for str_id in str_ids:
# id = int(str_id)
# if id >= 0:
# self.opt.gpu_ids.append(id)
args = vars(self.opt)
print('------------ Options -------------')
for k, v in sorted(args.items()):
print('%s: %s' % (str(k), str(v)))
print('-------------- End ----------------')
# save to the disk
expr_dir = os.path.join(self.opt.checkpoint_dir, self.name)
util.mkdirs(expr_dir)
file_name = os.path.join(expr_dir, 'opt.txt')
with open(file_name, 'wt') as opt_file:
opt_file.write('------------ Options -------------\n')
for k, v in sorted(args.items()):
opt_file.write('%s: %s\n' % (str(k), str(v)))
opt_file.write('-------------- End ----------------\n')
return self.opt
def __init__(self, opt):
self.opt = opt
self.use_html = opt.use_html
self.tf_log = opt.isTrain and not opt.use_html
self.win_size = opt.display_winsize
self.name = opt.name
self.log_dir = os.path.join(opt.checkpoints_dir, opt.name, 'logs')
util.mkdir(self.log_dir)
# if using tensorboard
if self.tf_log:
import tensorflow as tf
self.tf = tf
self.log_dir = os.path.join(opt.checkpoints_dir, opt.name, 'logs')
self.writer = tf.summary.FileWriter(self.log_dir)
# if using simple html page
if self.use_html:
self.web_dir = os.path.join(opt.checkpoints_dir, opt.name, 'web')
self.img_dir = os.path.join(self.web_dir, 'images')
util.mkdirs([self.web_dir, self.img_dir])
# save test image results
if not opt.isTrain:
self.test_dir = os.path.join(opt.checkpoints_dir, opt.name,
'test_img_out')
util.mkdirs([self.test_dir])
# log txt file head
self.log_name_txt = os.path.join(opt.checkpoints_dir, opt.name,
'loss_log.txt')
with open(self.log_name_txt, "a") as log_file:
now = time.strftime("%c")
log_file.write(
'================ Training Loss (%s) ================\n' % now)
# log csv file head
header = ['epoch', 'iters', 'time', 'loss_G', 'loss_D']
self.log_name_csv = os.path.join(opt.checkpoints_dir, opt.name,
'loss_log.csv')
with open(self.log_name_csv, "w") as log_file:
writer = csv.writer(log_file, delimiter=',')
writer.writerow(header)
# log csv file head for individual loss
self.log_individual_csv = os.path.join(opt.checkpoints_dir, opt.name,
'loss_individual_log.csv')
with open(self.log_individual_csv, "w") as log_file:
writer = csv.writer(log_file, delimiter=',')
writer.writerow(header)
# save loss graph as png
self.error_plot = os.path.join(opt.checkpoints_dir, opt.name,
'error_plot.png')
self.individual_error_plot = os.path.join(opt.checkpoints_dir,
opt.name,
'individual_error_plot.png')
def resume_logger(opt):
# parser = argparse.ArgumentParser()
# parser.add_argument(
# '-opt', type=str, required=True, help='Path to option JSON file.')
# opt = option.parse(json_path, is_train=True)
# opt = option.dict_to_nonedict(
# opt) # Convert to NoneDict, which return None for missing key.
# train from scratch OR resume training
if opt['path']['resume_state']: # resuming training
resume_state = torch.load(opt['path']['resume_state'])
else: # training from scratch
resume_state = None
util.mkdir_and_rename(
opt['path']['experiments_root']) # rename old folder if exists
util.mkdirs((path for key, path in opt['path'].items()
if not key == 'experiments_root'
and 'pretrain_model' not in key and 'resume' not in key))
# config loggers. Before it, the log will not work
util.setup_logger(None,
opt['path']['log'],
'train',
level=logging.INFO,
screen=True)
util.setup_logger('val', opt['path']['log'], 'val', level=logging.INFO)
logger = logging.getLogger('base')
return resume_state, logger
def parse(self):
if not self.initialized:
self.initialize()
self.opt = self.parser.parse_args()
# str_ids = self.opt.gpu_ids.split(',')
# self.opt.gpu_ids = []
# for str_id in str_ids:
# id = int(str_id)
# if id >= 0:
# self.opt.gpu_ids.append(id)
args = vars(self.opt)
print('------------ Options -------------')
for k, v in sorted(args.items()):
print('%s: %s' % (str(k), str(v)))
print('-------------- End ----------------')
# save to the disk
if self.opt.exp_id == '':
print('Please set the experimental ID with option --exp_id')
exit()
exp_dir = os.path.join(self.opt.exp_dir, self.opt.exp_id)
util.mkdirs(exp_dir)
if self.opt.resume_prefix_pose != '':
trunc_index = self.opt.resume_prefix_pose.index('pth')
self.opt.resume_prefix_pose = self.opt.resume_prefix_pose[
0:trunc_index - 1]
self.opt.resume_prefix_pose += '-'
# opt_name = self.opt.resume_prefix_pose + 'opt.txt'
# opt_name = os.path.join(exp_dir, opt_name)
# else:
# opt_name = os.path.join(exp_dir, 'opt.txt')
if self.opt.resume_prefix_asn != '':
trunc_index = self.opt.resume_prefix_asn.index('pth')
self.opt.resume_prefix_asn = self.opt.resume_prefix_asn[
0:trunc_index - 1]
self.opt.resume_prefix_asn += '-'
# opt_name = self.opt.resume_prefix_asdn + 'opt.txt'
# opt_name = os.path.join(exp_dir, opt_name)
# else:
# opt_name = os.path.join(exp_dir, 'opt.txt')
if self.opt.resume_prefix_dropout != '':
trunc_index = self.opt.resume_prefix_dropout.index('pth')
self.opt.resume_prefix_dropout = self.opt.resume_prefix_dropout[
0:trunc_index - 1]
self.opt.resume_prefix_dropout += '-'
if self.opt.resume_prefix_aug != '':
trunc_index = self.opt.resume_prefix_aug.index('pth')
self.opt.resume_prefix_aug = self.opt.resume_prefix_aug[
0:trunc_index - 1]
self.opt.resume_prefix_aug += '-'
with open('opt.txt', 'wt') as opt_file:
opt_file.write('------------ Options -------------\n')
for k, v in sorted(args.items()):
opt_file.write('%s: %s\n' % (str(k), str(v)))
opt_file.write('-------------- End ----------------\n')
return self.opt
def train(opt):
# Prepare the training corpus
print(options.TrainLogPrefix + "Prepare the training corpus begin!")
from datasource.input_corpus import InputCorpus
input_corpus = InputCorpus(opt.corpus_root, encoding=opt.encoding)
print(options.TrainLogPrefix + "Prepare the training corpus end!")
# Get the basic tfidf features
print(options.TrainLogPrefix + "Get the basic tfidf features begin!")
from feature.ngram_tfidf import NgramTfidf
ngram_tfidf = NgramTfidf(input_corpus)
ngram_tfidf.set_stopwords('./resource/stop_words_zh.utf8.txt')
import numpy as np
tfidf_mat, features = ngram_tfidf.get_tfidf_mat(top_k=opt.tfidf_top_k)
tfidf_mat = np.asarray(tfidf_mat)
features = np.asarray(features)
targets = np.asarray(input_corpus.get_filenames_and_targets()[1])
print(options.TrainLogPrefix + "Get the basic tfidf features end!")
# Do feature selection
print(options.TrainLogPrefix + "Do feature selection begin!")
if opt.which_filter == 'mi':
from feature.feature_selection import MISelection as FeatureSelection
feature_selector = FeatureSelection(tfidf_mat,
targets,
mi_threshold=opt.mi_threshold)
else:
from feature.feature_selection import GBDTSelection as FeatureSelection
feature_selector = FeatureSelection(tfidf_mat, targets)
boolean_selection_index = feature_selector.get_boolean_selection_lst()
filtered_tfidf_mat = tfidf_mat[:, boolean_selection_index]
filtered_features = features[boolean_selection_index]
print(options.TrainLogPrefix + "Do feature selection end!")
# Training model
print(options.TrainLogPrefix + "Training model begin!")
if opt.which_classifier == 'svm':
from model.classifier import SVMClassifier as Classifier
else:
from model.classifier import GBDTClassifier as Classifier
classifier_model = Classifier()
from model.classifier import Scorer
scorer = Scorer(classifier_model.get_model(), filtered_tfidf_mat, targets)
print(options.TrainLogPrefix + "Training model end!")
scorer.show_score()
# Save the model
model_save_path = opt.path_to_save_model
from utils import util
util.mkdirs('/'.join(model_save_path.split('/')[:-1]))
classifier_model.dump(filtered_tfidf_mat, targets, model_save_path)
print(options.TrainLogPrefix + 'model save to ' + model_save_path)
# Save the filtered features
filtered_features_save_path = opt.path_to_save_model + options.FeaturesSaveSuffix
df_vec = ngram_tfidf.numDocsContainingFeatures(filtered_features)
save_features_df(df_vec, filtered_features, len(tfidf_mat),
filtered_features_save_path)
def _save(self, args):
expr_dir = os.path.join(self._opt.checkpoints_dir, self._opt.name)
print(expr_dir)
util.mkdirs(expr_dir)
file_name = os.path.join(expr_dir, 'opt_%s.txt' % ('train' if self.is_train else 'test'))
with open(file_name, 'wt') as opt_file:
opt_file.write('------------ Options -------------\n')
for k, v in sorted(args.items()):
opt_file.write('%s: %s\n' % (str(k), str(v)))
opt_file.write('-------------- End ----------------\n')
def _save(self, args):
expr_dir = os.path.join(self._opt.checkpoints_dir, self._opt.name)
print(expr_dir)
util.mkdirs(expr_dir)
file_name = os.path.join(expr_dir, 'opt_%s.txt' % ('train' if self.is_train else 'test'))
with open(file_name, 'wt') as opt_file:
opt_file.write('------------ Options -------------\n')
for k, v in sorted(args.items()):
opt_file.write('%s: %s\n' % (str(k), str(v)))
opt_file.write('-------------- End ----------------\n')
def initialize(self, opt):
self.opt = opt
self.training = opt.isTrain
self.gpu_ids = opt.gpu_ids
self.isTrain = opt.isTrain
self.num_classes = opt.label_nc
self.Tensor = torch.cuda.FloatTensor if self.gpu_ids else torch.Tensor
self.save_dir = os.path.join(opt.checkpoints_dir, opt.name)
self.tensorborad_dir = os.path.join(self.opt.checkpoints_dir, 'tensorboard', opt.dataset_mode)
self.model_dir = os.path.join(self.opt.checkpoints_dir, self.opt.name, 'model')
util.mkdirs([self.tensorborad_dir, self.model_dir])
def save_models(self, epoch, save_dir, fid=None, isbest=False, direction='AtoB'):
util.mkdirs(save_dir)
ckpt = {
'G': self.netG.state_dict(),
'D': self.netD.state_dict(),
'epoch': epoch,
'fid': fid
}
if isbest:
torch.save(ckpt, os.path.join(save_dir, 'model_best_%s.pth' % direction))
else:
torch.save(ckpt, os.path.join(save_dir, 'model_%d.pth' % epoch))
def adaptive_personalize(opt, imitator, visualizer, output_path):
output_dir = opt.output_dir
mkdirs([os.path.join(output_dir, 'imgs'), os.path.join(output_dir, 'pairs')])
# TODO check if it has been computed.
print('\n\t\t\tPersonalization: meta imitation...')
imitator.personalize(opt.src_path, visualizer=None, output_path = output_path)
meta_imitate(opt, imitator, prior_tgt_path=opt.pri_path, visualizer=None, save_imgs=True)
# post tune
print('\n\t\t\tPersonalization: meta cycle finetune...')
loader = make_dataset(opt)
imitator.post_personalize(opt.output_dir, loader, visualizer=None, verbose=False)
def dir_check(opt):
if opt['is_train']:
# starting from scratch, needs to create training directory
if not opt['path']['resume_state']:
util.mkdir_and_rename(
opt['path']['experiments_root']) # rename old folder if exists
util.mkdirs(
(path for key, path in opt['path'].items()
if not key == 'experiments_root'
and 'pretrain_model' not in key and 'resume' not in key))
else:
# create testing directory
util.mkdirs((path for key, path in opt['path'].items()
if not key == 'pretrain_model_G'))
def save_networks(self, epoch):
mkdirs(self.save_dir)
for name in self.model_names:
if isinstance(name, str):
save_filename = '%s_net_%s.pth' % (epoch, name)
save_path = os.path.join(self.save_dir, save_filename)
net = getattr(self, 'net' + name)
if torch.cuda.is_available():
torch.save(net.cpu().state_dict(), save_path)
net.to(self.device)
else:
torch.save(net.cpu().state_dict(), save_path)
def test_pix2pix_mIoU(model, opt):
opt.phase = 'val'
opt.num_threads = 0
opt.batch_size = 1
opt.serial_batches = True
opt.no_flip = True
opt.load_size = 256
opt.display_id = -1
dataset = create_dataset(opt)
model.model_eval()
result_dir = os.path.join(opt.checkpoints_dir, opt.name, 'test_results')
util.mkdirs(result_dir)
fake_B = {}
names = []
for i, data in enumerate(dataset):
model.set_input(data)
with torch.no_grad():
model.forward()
visuals = model.get_current_visuals()
fake_B[data['A_paths'][0]] = visuals['fake_B']
for path in range(len(model.image_paths)):
short_path = ntpath.basename(model.image_paths[0][0])
name = os.path.splitext(short_path)[0]
if name not in names:
names.append(name)
util.save_images(visuals,
model.image_paths,
result_dir,
direction=opt.direction,
aspect_ratio=opt.aspect_ratio)
drn_model = DRNSeg('drn_d_105', 19, pretrained=False).to(model.device)
util.load_network(drn_model, opt.drn_path, verbose=False)
drn_model.eval()
mIoU = get_mIoU(list(fake_B.values()),
names,
drn_model,
model.device,
table_path=os.path.join(opt.dataroot, 'table.txt'),
data_dir=opt.dataroot,
batch_size=opt.batch_size,
num_workers=opt.num_threads)
return mIoU
def initialize(self, opt):
TIMESTAMP = "{0:%Y-%m-%dT%H-%M-%S/}".format(datetime.now())
self.opt = opt
self.training = opt.isTrain
self.gpu_ids = opt.gpu_ids
self.isTrain = opt.isTrain
self.num_classes = opt.label_nc
self.Tensor = torch.cuda.FloatTensor if self.gpu_ids else torch.Tensor
self.save_dir = os.path.join(opt.checkpoints_dir, opt.name)
self.tensorborad_dir = os.path.join(self.opt.checkpoints_dir,
'tensorboard',
opt.dataset_mode + TIMESTAMP)
self.model_dir = os.path.join(self.opt.checkpoints_dir, self.opt.name,
'model')
util.mkdirs([self.tensorborad_dir, self.model_dir])
def save_models(self, epoch, save_dir, fid=None, isbest=False, direction='AtoB'):
util.mkdirs(save_dir)
ckpt = {
'G_A': self.__pop_ops_params_state_dict(self.netG_A.state_dict()),
'G_B': self.__pop_ops_params_state_dict(self.netG_B.state_dict()),
'D_A': self.netD_A.state_dict(),
'D_B': self.netD_B.state_dict(),
'epoch': epoch,
'cfg': (self.cfg_AtoB, self.cfg_BtoA),
'fid': fid
}
if isbest:
torch.save(ckpt, os.path.join(save_dir, 'model_best_%s.pth' % direction))
else:
torch.save(ckpt, os.path.join(save_dir, 'model_%d.pth' % epoch))
def save_models(self, model, epoch, name, opt, isbest):
save_dir = os.path.join(opt.checkpoints_dir, opt.name, 'checkpoints')
utils.mkdirs(save_dir)
ckpt = {
'weight': model.module.state_dict(),
'epoch': epoch,
'cfg': opt.model,
'index': name
}
if isbest:
torch.save(ckpt, os.path.join(save_dir, 'model%s_best.pth' % name))
else:
torch.save(
ckpt, os.path.join(save_dir, 'model%s_%d.pth' % (name, epoch)))
def meta_imitate(opt, imitator, prior_tgt_path, save_imgs=True, visualizer=None):
src_path = opt.src_path
all_tgt_paths = scan_tgt_paths(prior_tgt_path, itv=40)
output_dir = opt.output_dir
out_img_dir, out_pair_dir = mkdirs([os.path.join(output_dir, 'imgs'), os.path.join(output_dir, 'pairs')])
img_pair_list = []
for t in tqdm(range(len(all_tgt_paths))):
tgt_path = all_tgt_paths[t]
preds = imitator.inference([tgt_path], visualizer=visualizer, cam_strategy=opt.cam_strategy, verbose=False)
tgt_name = os.path.split(tgt_path)[-1]
out_path = os.path.join(out_img_dir, 'pred_' + tgt_name)
if save_imgs:
cv_utils.save_cv2_img(preds[0], out_path, normalize=True)
write_pair_info(imitator.src_info, imitator.tsf_info,
os.path.join(out_pair_dir, '{:0>8}.pkl'.format(t)), imitator=imitator,
only_vis=opt.only_vis)
img_pair_list.append((src_path, tgt_path))
if save_imgs:
write_pickle_file(os.path.join(output_dir, 'pairs_meta.pkl'), img_pair_list)
def parse(self):
if not self.initialized:
self.initialize()
self.opt = self.parser.parse_args()
self.opt.isTrain = self.isTrain # train or test
args = vars(self.opt)
if self.opt.dataset_mode == 'sample_per_vehicle':
self.opt.frames_per_sample = max(self.opt.n_rgbs_per_sample,
self.opt.n_bbs_per_sample)
self.opt.input_channel_dim = 3 * self.opt.n_rgbs_per_sample + self.opt.n_bbs_per_sample
self.opt.train_frames_to_remove = int(
self.opt.train_frames_per_scene - self.opt.ttc_threshold * 10 -
self.opt.frames_per_sample + 1)
self.opt.valid_frames_to_remove = int(
self.opt.valid_frames_per_scene - self.opt.ttc_threshold * 10 -
self.opt.frames_per_sample + 1)
self.opt.test_frames_to_remove = int(
self.opt.test_frames_per_scene - self.opt.ttc_threshold * 10 -
self.opt.frames_per_sample + 1)
elif self.opt.dataset_mode == 'sample_per_frame':
self.opt.input_channel_dim = 26
else:
print("Unknown dataset mode!")
raise ValueError
if self.opt.shuffle_by_which not in ['none', 'sample', 'frame']:
print("Unknown shuffling method")
raise ValueError
print('------------ Options -------------')
for k, v in sorted(args.items()):
print('%s: %s' % (str(k), str(v)))
print('-------------- End ----------------')
if self.isTrain:
# save to the disk
expr_dir = os.path.join(self.opt.checkpoints_dir, self.opt.name)
util.mkdirs(expr_dir)
file_name = os.path.join(expr_dir, 'opt.txt')
with open(file_name, 'wt') as opt_file:
opt_file.write('------------ Options -------------\n')
for k, v in sorted(args.items()):
opt_file.write('%s: %s\n' % (str(k), str(v)))
opt_file.write('-------------- End ----------------\n')
return self.opt
def __init__(self, opt):
# self.opt = opt
self.tf_log = opt.tf_log
self.use_html = opt.isTrain and not opt.no_html
self.win_size = opt.display_winsize
self.name = opt.name
if self.tf_log:
import tensorflow as tf
self.tf = tf
self.log_dir = os.path.join(opt.checkpoints_dir, opt.name, 'logs')
self.writer = tf.summary.FileWriter(self.log_dir)
if self.use_html:
self.web_dir = os.path.join(opt.checkpoints_dir, opt.name, 'web')
self.img_dir = os.path.join(self.web_dir, 'images')
print('create web directory %s...' % self.web_dir)
util.mkdirs([self.web_dir, self.img_dir])
def setup_logging(opt, resume_state, rank):
tb_logger = None
if rank <= 0: # normal training (rank -1) OR distributed training (rank 0)
if resume_state is None:
util.mkdir_and_rename(
opt['path']
['experiments_root']) # rename experiment folder if exists
util.mkdirs(
(path for key, path in opt['path'].items()
if not key == 'experiments_root'
and 'pretrain_model' not in key and 'resume' not in key))
# config loggers. Before it, the log will not work
util.setup_logger('base',
opt['path']['log'],
'train_' + opt['name'],
level=logging.INFO,
screen=True,
tofile=True)
util.setup_logger('val',
opt['path']['log'],
'val_' + opt['name'],
level=logging.INFO,
screen=True,
tofile=True)
logger = logging.getLogger('base')
logger.info(option.dict2str(opt))
# tensorboard logger
if opt['use_tb_logger'] and 'debug' not in opt['name']:
version = float(torch.__version__[0:3])
if version >= 1.1: # PyTorch 1.1
from torch.utils.tensorboard import SummaryWriter
else:
logger.info(
'You are using PyTorch {}. Tensorboard will use [tensorboardX]'
.format(version))
from tensorboardX import SummaryWriter
tb_logger = SummaryWriter(log_dir='../tb_logger/' + opt['name'])
else:
util.setup_logger('base',
opt['path']['log'],
'train',
level=logging.INFO,
screen=True)
logger = logging.getLogger('base')
return logger, tb_logger
def check_args(args):
args.isTrain = args.phase == 'train'
str_ids = args.gpu_ids.split(',')
args.gpu_ids = []
for str_id in str_ids:
id = int(str_id)
if id >= 0:
args.gpu_ids.append(id)
if len(args.gpu_ids) > 0:
torch.cuda.set_device(args.gpu_ids[0])
try:
assert args.batch_size >= 1
except:
print('batch size must be larger than or equal to one')
expr_dir = os.path.join(args.checkpoints_dir, args.name)
util.mkdirs(expr_dir)
return args
def save(self, args):
expr_dir = os.path.join(self.opt.checkpoints_dir, self.opt.name)
print(expr_dir)
util.mkdirs(expr_dir)
file_name = os.path.join(expr_dir, 'opt_%s.txt' %
('train' if self.is_train else 'test'))
file_name_yaml = os.path.join(expr_dir, 'opt_%s.yaml' %
('train' if self.is_train else 'test'))
with open(file_name_yaml, 'w') as opt_file:
yaml.dump(args, opt_file)
with open(file_name, 'wt') as opt_file:
opt_file.write('------------ Options -------------\n')
for k, v in sorted(args.items()):
opt_file.write('%s: %s\n' % (str(k), str(v)))
opt_file.write('-------------- End ----------------\n')
file_name = os.path.join(expr_dir, 'command_line.txt')
with open(file_name, 'wt') as opt_file:
opt_file.write(" ".join(sys.argv))
def save_models(self,
epoch,
save_dir,
fid=None,
isbest=False,
direction='AtoB'):
util.mkdirs(save_dir)
ckpt = {
'G': self.__pop_ops_params_state_dict(self.netG.state_dict()),
'D': self.netD.state_dict(),
'epoch': epoch,
'cfg': (self.filter_cfgs, self.channel_cfgs),
'fid': fid
}
if isbest:
torch.save(ckpt,
os.path.join(save_dir, 'model_best_%s.pth' % direction))
else:
torch.save(ckpt, os.path.join(save_dir, 'model_%d.pth' % epoch))
def print_options(self, opt):
message = ''
message += '----------------- Options ---------------\n'
for k, v in sorted(vars(opt).items()):
comment = ''
default = self.parser.get_default(k)
if v != default:
comment = '\t[default: %s]' % str(default)
message += '{:>25}: {:<30}{}\n'.format(str(k), str(v), comment)
message += '----------------- End -------------------'
print(message)
# save to the disk
expr_dir = os.path.join(opt.checkpoints_dir, opt.name)
util.mkdirs(expr_dir)
file_name = os.path.join(expr_dir, 'opt.txt')
with open(file_name, 'wt') as opt_file:
opt_file.write(message)
opt_file.write('\n')
def save_img_metric(self, img_dict, path, model_name, phase):
util.mkdirs(os.path.join(path, model_name, phase, 'A2B', 'depth'))
# util.mkdirs(os.path.join(path, model_name, phase, 'A2B', 'normal'))
util.mkdirs(os.path.join(path, model_name, phase, 'B2A', 'depth'))
# util.mkdirs(os.path.join(path, model_name, phase, 'B2A', 'normal'))
# util.mkdirs(os.path.join(path, model_name, phase, 'A2B2A', 'depth'))
# util.mkdirs(os.path.join(path, model_name, phase, 'A2B2A', 'normal'))
# util.mkdirs(os.path.join(path, model_name, phase, 'B2A2B', 'depth'))
# util.mkdirs(os.path.join(path, model_name, phase, 'B2A2B', 'normal'))
B_depth_fake = util.tensor2mm(img_dict['fake_depth_B'], self.opt)
A_name = img_dict['name_A']
A_depth_fake = util.tensor2mm(img_dict['fake_depth_A'], self.opt)
B_name = img_dict['name_B']
for i in range(B_depth_fake.shape[0]):
imageio.imwrite(os.path.join(path, model_name, phase, 'A2B', 'depth', A_name[i]+'.png'), B_depth_fake[i])
# imageio.imwrite(os.path.join(path, model_name, phase, 'A2B2A', 'depth', A_name[i]+'.png'), A_rec[i].astype(np.uint16))
# np.save(os.path.join(path, model_name, phase, 'A2B', 'normal', A_name[i]+'.npy'), util.get_normal_metric(B_depth_fake[i]))
imageio.imwrite(os.path.join(path, model_name, phase, 'B2A', 'depth', B_name[i]+'.png'), A_depth_fake[i])
def save_img(self, img_dict, path, model_name, phase):
util.mkdirs(os.path.join(path, model_name, phase, 'A'))
util.mkdirs(os.path.join(path, model_name, phase, 'B'))
A_imgs = util.tensor2im(img_dict['real_img_A'], self.opt, isDepth=False)
A_depth = util.tensor2im(img_dict['real_depth_A'], self.opt, isDepth=True)*1000
B_depth_fake = util.tensor2im(img_dict['fake_depth_B'], self.opt, isDepth=True)*1000
A_name = img_dict['name_A']
B_imgs = util.tensor2im(img_dict['real_img_B'], self.opt, isDepth=False)
B_depth = util.tensor2im(img_dict['real_depth_B'], self.opt, isDepth=True)*1000
A_depth_fake = util.tensor2im(img_dict['fake_depth_A'], self.opt, isDepth=True)*1000
B_name = img_dict['name_B']
for i in range(A_imgs.shape[0]):
imageio.imwrite(os.path.join(path, model_name, phase, 'A', A_name[i]+'_img.png'), A_imgs[i])
imageio.imwrite(os.path.join(path, model_name, phase, 'A', A_name[i]+'_depth.png'), A_depth[i].astype(np.uint16))
imageio.imwrite(os.path.join(path, model_name, phase, 'A', A_name[i]+'_depth_fake.png'), B_depth_fake[i].astype(np.uint16))
imageio.imwrite(os.path.join(path, model_name, phase, 'B', B_name[i]+'_img.png'), B_imgs[i])
imageio.imwrite(os.path.join(path, model_name, phase, 'B', B_name[i]+'_depth.png'), B_depth[i].astype(np.uint16))
imageio.imwrite(os.path.join(path, model_name, phase, 'B', B_name[i]+'_depth_fake.png'), A_depth_fake[i].astype(np.uint16))
def get_options(json_path):
"""options"""
# parser = argparse.ArgumentParser()
# parser.add_argument(
# '-opt', type=str, required=True, help='Path to options JSON file.')
# opt = option.parse(parser.parse_args().opt, is_train=False)
is_train = False
opt = option.parse(json_path, is_train)
util.mkdirs((path for key, path in opt['path'].items()
if not key == 'pretrain_model_G'))
opt = option.dict_to_nonedict(opt)
util.setup_logger(None,
opt['path']['log'],
'test',
level=logging.INFO,
screen=True)
logger = logging.getLogger('base')
logger.info(option.dict2str(opt))
return opt, logger
def print_options(opt):
"""print and save options"""
# print('--------------Options--------------')
# for k, v in sorted(vars(opt).items()):
# print('%s: %s' % (str(k), str(v)))
# print('----------------End----------------')
# save to the disk
expr_dir = os.path.join(opt.checkpoints_dir, opt.name)
util.mkdirs(expr_dir)
if opt.isTrain:
file_name = os.path.join(expr_dir, 'train_opt.txt')
else:
file_name = os.path.join(expr_dir, 'test_opt.txt')
with open(file_name, 'wt') as opt_file:
opt_file.write('--------------Options--------------\n')
for k, v in sorted(vars(opt).items()):
opt_file.write('%s: %s\n' % (str(k), str(v)))
opt_file.write('----------------End----------------\n')
