|
|
@ -214,7 +214,8 @@ class FakeDataReader(object):
|
|
|
|
self.img_std = np.array(cfg.MODEL.image_std).reshape(
|
|
|
|
self.img_std = np.array(cfg.MODEL.image_std).reshape(
|
|
|
|
[3, 1, 1]).astype(np.float32)
|
|
|
|
[3, 1, 1]).astype(np.float32)
|
|
|
|
|
|
|
|
|
|
|
|
self.batch_size = cfg[mode.upper()]['batch_size']
|
|
|
|
self.batch_size = 1 if sys.platform == 'darwin' or os.name == 'nt' else cfg[
|
|
|
|
|
|
|
|
mode.upper()]['batch_size']
|
|
|
|
self.generator_out = []
|
|
|
|
self.generator_out = []
|
|
|
|
self.total_iter = 3
|
|
|
|
self.total_iter = 3
|
|
|
|
for i in range(self.total_iter):
|
|
|
|
for i in range(self.total_iter):
|
|
|
@ -240,7 +241,8 @@ class FakeDataReader(object):
|
|
|
|
|
|
|
|
|
|
|
|
def create_optimizer(cfg, params):
|
|
|
|
def create_optimizer(cfg, params):
|
|
|
|
total_videos = cfg.total_videos
|
|
|
|
total_videos = cfg.total_videos
|
|
|
|
step = int(total_videos / cfg.batch_size + 1)
|
|
|
|
batch_size = 1 if sys.platform == 'darwin' or os.name == 'nt' else cfg.batch_size
|
|
|
|
|
|
|
|
step = int(total_videos / batch_size + 1)
|
|
|
|
bd = [e * step for e in cfg.decay_epochs]
|
|
|
|
bd = [e * step for e in cfg.decay_epochs]
|
|
|
|
base_lr = cfg.learning_rate
|
|
|
|
base_lr = cfg.learning_rate
|
|
|
|
lr_decay = cfg.learning_rate_decay
|
|
|
|
lr_decay = cfg.learning_rate_decay
|
|
|
|