def save(self, save_optimizer=False, save_path=None, **kwargs):
"""serialize models include optimizer and other info
return path where the model-file is stored.
Args:
save_optimizer (bool): whether save optimizer.state_dict().
save_path (string): where to save model, if it's None, save_path
is generate using time str and info from kwargs.
Returns:
save_path(str): the path to save models.
"""
save_dict = dict()
save_dict['model'] = self.faster_rcnn.state_dict()
save_dict['config'] = opt._state_dict()
save_dict['other_info'] = kwargs
save_dict['vis_info'] = self.vis.state_dict()
if save_optimizer:
save_dict['optimizer'] = self.optimizer.state_dict()
if save_path is None:
timestr = time.strftime('%m%d%H%M')
save_path = 'checkpoints/fasterrcnn_%s' % timestr
for k_, v_ in kwargs.items():
save_path += '_%s' % v_
t.save(save_dict, save_path)
self.vis.save([self.vis.env])
return save_path
def save(self,
save_optimizer=False,
save_path=None,
prune=False,
**kwargs):
"""serialize models include optimizer and other info
return path where the model-file is stored.
Args:
save_optimizer (bool): whether save optimizer.state_dict().
save_path (string): where to save model, if it's None, save_path
is generate using time str and info from kwargs.
Returns:
save_path(str): the path to save models.
"""
save_dict = dict()
save_dict['sparse_list'] = []
if self.sparse:
for n, m in self.named_modules():
if hasattr(m, "sparse"):
if m.sparse and hasattr(m, 'weight'):
w_dev = m.weight.device
w = m.weight.data.coalesce().to_dense()
m.weight.data = w.to(w_dev)
save_dict['sparse_list'].append(str(m))
save_dict['model'] = self.faster_rcnn.state_dict()
save_dict['config'] = opt._state_dict()
save_dict['other_info'] = kwargs
save_dict['vis_info'] = self.vis.state_dict()
save_dict['sparse'] = self.sparse
if save_optimizer:
save_dict['optimizer'] = self.optimizer.state_dict()
if save_path is None:
timestr = time.strftime('%m%d%H%M')
save_path = 'checkpoints/fasterrcnn_%s' % timestr
for k_, v_ in kwargs.items():
save_path += '_%s' % v_
if prune:
save_path += "_prune"
save_dir = os.path.dirname(save_path)
if not os.path.exists(save_dir):
os.makedirs(save_dir)
t.save(save_dict, save_path)
self.vis.save([self.vis.env])
return save_path
def save(self, save_optimizer=False, save_path=None, **kwargs):
"""serialize models include optimizer and other info
return path where the model-file is stored.
Args:
save_optimizer (bool): whether save optimizer.state_dict().
save_path (string): where to save model, if it's None, save_path
is generate using time str and info from kwargs.
Returns:
save_path(str): the path to save models.
"""
save_dict = dict()
save_dict['model'] = self.faster_rcnn.state_dict()
save_dict['config'] = opt._state_dict()
save_dict['other_info'] = kwargs
save_dict['vis_info'] = self.vis.state_dict()
if save_optimizer:
save_dict['optimizer'] = self.optimizer.state_dict()
if save_path is None:
# timestr = time.strftime('%m%d%H%M')
# save_path = 'checkpoints/fasterrcnn_%s' % timestr
# for k_, v_ in kwargs.items():
# save_path += '_%s' % v_
save_path = os.path.join(opt.logs_path, opt.model_name, 'results')
os.makedirs(save_path, exist_ok=True)
# save_dir = os.path.dirname(save_path)
# if not os.path.exists(save_dir):
# os.makedirs(save_dir)
t.save(save_dict, os.path.join(save_path, opt.model_name + '.pt'))
self.vis.save([self.vis.env])
with open(os.path.join(save_path, 'infos.json'), 'w') as fp:
json.dump(save_dict['other_info'], fp)
return save_path
def save(self,
save_optimizer=False,
save_path=None,
save_rcnn=True,
**kwargs):
"""serialize models include optimizer and other info
return path where the model-file is stored.
Args:
save_optimizer (bool): whether save optimizer.state_dict().
save_path (string): where to save model, if it's None, save_path
is generate using time str and info from kwargs.
Returns:
save_path(str): the path to save models.
"""
save_dict = dict()
save_dict['model'] = self.faster_rcnn.state_dict()
save_dict['config'] = opt._state_dict()
save_dict['other_info'] = kwargs
save_dict['vis_info'] = self.vis.state_dict()
if save_optimizer:
save_dict['optimizer'] = self.optimizer.state_dict()
if save_path is None:
timestr = time.strftime('%m%d%H%M')
save_path = 'checkpoints/fasterrcnn_full_%s' % timestr
if not self.attack_mode:
for k_, v_ in kwargs.items():
save_path += '_%s' % v_
if self.attacker is not None:
self.attacker.save('checkpoints/max_min_attack_6.pth')
if save_rcnn:
t.save(save_dict, save_path)
self.vis.save([self.vis.env])
return save_path
def save(self, save_optimizer=False, save_path=None, save_rcnn=False, **kwargs):
save_dict = dict()
save_dict['model'] = self.faster_rcnn.state_dict()
save_dict['config'] = opt._state_dict()
save_dict['other_info'] = kwargs
save_dict['vis_info'] = self.vis.state_dict()
if save_optimizer:
save_dict['optimizer'] = self.optimizer.state_dict()
if save_path is None:
timestr = time.strftime('%m%d%H%M')
save_path = 'checkpoints/faterrcnn_full_%s' % timestr
if not self.attack_mode:
for k_, v_ in kwargs.items():
save_path += '%s' % v_
if self.attacker is not None:
self.attacker.save('checkpoints/attack_%s_%d.path' % (timestr, kwargs['epochs']))
if save_rcnn:
t.save(save_dict, save_path)
self.vis.save([self.vis.env])
return save_path
def save(self, save_optimizer=False, save_path=None, **kwargs):
save_dict= dict()
save_dict['model'] = self.faster_rcnn.state_dict()
save_dict['config'] = opt._state_dict()
save_dict['other_info'] = kwargs
save_dict['vis_info'] = self.vis.state_dict()
if save_optimizer:
save_dict['optimizer'] = self.optimizer.state_dict()
if save_path is None:
timestr = time.strftime('%m%d%H%M')
save_path = 'checkpoints/fasterrcnn_%s' % timestr
for k_, v_ in kwargs.items():
save_path += '_%s' % v_
save_dir = os.path.dirname(save_path)
if not os.path.exists(save_dir):
os.makedirs(save_dir)
torch.save(save_dict, save_path)
self.vis.save([self.vis.env])
return save_path
def save(self, save_optimizer=False, save_path=None, **kwargs):
"""serialize models include optimizer and other info
return path where the model-file is stored.
Args:
save_optimizer (bool): whether save optimizer.state_dict().
save_path (string): where to save model, if it's None, save_path
is generate using time str and info from kwargs.
Returns:
save_path(str): the path to save models.
"""
save_dict = dict()
save_dict['model'] = self.faster_rcnn.state_dict()
save_dict['config'] = opt._state_dict()
save_dict['other_info'] = kwargs
if save_optimizer:
save_dict['optimizer'] = self.optimizer.state_dict()
if save_path is None:
timestr = time.strftime('%m%d%H%M')
import os
# import stat
save_path = os.getcwd() + '\\checkpoints\\fasterrcnn_%s' % timestr
# print('save_path:', save_path)
os.makedirs(save_path)
# os.chmod(save_path, stat.S_IRWXU|stat.S_IRWXG|stat.S_IRWXO)
# os.chmod(save_path, stat.S_IRWXO)
for k_, v_ in kwargs.items():
save_path += '_%s' % v_
t.save(save_dict, save_path) # permission denied 这里还要处理下
return save_path
def save(self, save_optimizer=False, save_path=None, **kwargs):
"""serialize models include optimizer and other info
return path where the model-file is stored.
Args:
save_optimizer (bool): whether save optimizer.state_dict().
save_path (string): where to save model, if it's None, save_path
is generate using time str and info from kwargs.
Returns:
save_path(str): the path to save models.
"""
save_dict = dict()
save_dict['model'] = self.faster_rcnn.state_dict()
save_dict['config'] = opt._state_dict()
save_dict['other_info'] = kwargs
save_dict['vis_info'] = self.vis.state_dict()
if save_optimizer:
save_dict['optimizer'] = self.optimizer.state_dict()
if save_path is None:
timestr = time.strftime('%m%d%H%M')
save_path = 'checkpoints/fasterrcnn_%s' % timestr
for k_, v_ in kwargs.items():
if k_ == 'best_map':
save_path += '_%s' % v_
break
save_dir = os.path.dirname(save_path)
if not os.path.exists(save_dir):
os.makedirs(save_dir)
t.save(save_dict, save_path)
self.vis.save([self.vis.env])
return save_path
os.mkdir(temp_path)
opt.model_dir = os.path.join(opt.model_sql_db, date,
datetime.now().strftime("%H:%M:%S"))
os.mkdir(opt.model_dir)
os.mkdir(os.path.join(opt.model_dir, 'figs'))
# Initialize Logger
base = '' if opt.mode != CONST.DEPLOY else '-' + os.path.basename(
opt.deploy_data).replace('.csv', '')
log_fname = '{}{}.log'.format(opt.mode, base)
Logger(log_filename=os.path.join(opt.model_dir, log_fname),
level=opt.logging_level,
log2file=opt.log2file)
logger = logging.getLogger('go-train')
Logger.section_break('User Config')
logger.info(pformat(opt._state_dict()))
# Initialize Tensorboard Logger
if opt.mode == CONST.TRAIN:
tb_logger = SummaryWriter(opt.model_dir)
else:
tb_logger = None
# Train and evaluate
if opt.mode == CONST.TRAIN or opt.mode == CONST.VAL:
train_and_evaluate(opt, logger, tb_logger)
else:
deploy(opt, logger)
if opt.save_model_db or sql_yn == 'y':
logger.info('Inserting current model to sql database')
