def _process_model_folder(path):
try:
model_configs[key] = load_json(str(path / "config.json"))
try:
# not all models have associated thresholds
model_threshs[key] = load_json(
str(path / "thresholds.json"))
except FileNotFoundError:
pass
finally:
select_model(key)
plugin.progress_bar.hide()
def _model_folder_change(_path: str):
path = Path(_path)
key = CUSTOM_MODEL, path
try:
if not path.is_dir():
return
model_configs[key] = load_json(str(path / "config.json"))
model_threshs[key] = load_json(str(path / "thresholds.json"))
except FileNotFoundError:
pass
finally:
select_model(key)
def __init__(self, config, name=None, basedir='.'):
super().__init__(config=config, name=name, basedir=basedir)
threshs = dict(prob=None, nms=None)
if basedir is not None:
try:
threshs = load_json(str(self.logdir / 'thresholds.json'))
print("Loading thresholds from 'thresholds.json'.")
if threshs.get('prob') is None or not (0 < threshs.get('prob')
< 1):
print(
"- Invalid 'prob' threshold (%s), using default value."
% str(threshs.get('prob')))
threshs['prob'] = None
if threshs.get('nms') is None or not (0 < threshs.get('nms') <
1):
print(
"- Invalid 'nms' threshold (%s), using default value."
% str(threshs.get('nms')))
threshs['nms'] = None
except FileNotFoundError:
if config is None and len(tuple(self.logdir.glob('*.h5'))) > 0:
print(
"Couldn't load thresholds from 'thresholds.json', using default values. "
"(Call 'optimize_thresholds' to change that.)")
self.thresholds = dict(
prob=0.5 if threshs['prob'] is None else threshs['prob'],
nms=0.4 if threshs['nms'] is None else threshs['nms'],
)
print(
"Using default values: prob_thresh={prob:g}, nms_thresh={nms:g}.".
format(prob=self.thresholds.prob, nms=self.thresholds.nms))
def _set_logdir(self):
if self.name is None:
self.name = datetime.datetime.now().strftime(
"%Y-%m-%d-%H-%M-%S.%f")
self.logdir = self.basedir / self.name
config_file = self.logdir / 'config.json'
if self.config is None:
if config_file.exists():
config_dict = load_json(str(config_file))
self.config = Config(**config_dict)
if not self.config.is_valid():
invalid_attr = self.config.is_valid(True)[1]
raise ValueError('Invalid attributes in loaded config: ' +
', '.join(invalid_attr))
else:
raise FileNotFoundError("config file doesn't exist: %s" %
str(config_file.resolve()))
else:
if self.logdir.exists():
warnings.warn(
'output path for model already exists, files may be overwritten: %s'
% str(self.logdir.resolve()))
self.logdir.mkdir(parents=True, exist_ok=True)
save_json(vars(self.config), str(config_file))
def set_bioseg_model(basedir,
model_name,
BioConfig,
load_tf_file='',
load_model=False,
seed=0):
mkdir(basedir)
print('------------------ BioSeg setup -------------------------')
print('basedir/model name : ', basedir + model_name)
print('model file : ',
basedir + model_name + '/' + BioConfig.train_checkpoint)
print()
if seed > 0:
tf.set_random_seed(seed)
if load_model:
print('** Loading Model Config **')
print()
if os.path.exists(basedir + model_name + '/config.json'):
print('Loading previous config file : ')
config_dict = load_json(basedir + model_name + '/config.json')
BioConfig = BioSegConfig(np.array([]), **config_dict)
print(basedir + model_name + '/config.json')
if load_tf_file == '': #load default best model in folder
load_tf_file = basedir + model_name + '/' + BioConfig.train_checkpoint
print('** Create Model **')
print()
print(BioConfig)
model = BioSeg_standard(BioConfig, model_name, basedir=basedir)
model.prepare_for_training()
if os.path.exists(load_tf_file):
print('Loading model weights : ', load_tf_file)
model.keras_model.load_weights(load_tf_file)
else:
print('Model with random initialization')
return model
def load(path):
return Config(**load_json(path))
