def get_student_cfg(cfg, path):
if os.path.exists(path):
with open(path) as file:
student_file_cfg = config.load_cfg(file)
else:
print("File {} not exists".format(args.student_file))
print("Please input a specific student config file")
print("Exiting...")
exit(-1)
student_cfg = copy.deepcopy(cfg)
for key, value in student_file_cfg.items():
student_cfg[key] = value
return student_cfg
def load_cfg_from_file(cfg_filename):
"""Load config from a file
Args:
cfg_filename (str):
Returns:
CfgNode: loaded configuration
"""
with open(cfg_filename, "r") as f:
cfg = load_cfg(f)
cfg_template = _C
cfg_template.merge_from_other_cfg(cfg)
return cfg_template
def make_config(config_file):
with open(config_file, 'r') as f:
cfg = config.load_cfg(f)
cfg.set_new_allowed(True)
return cfg.clone()
import numpy as np
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument("--mat", required=True)
parser.add_argument("--weights", required=True)
parser.add_argument("--var_name", required=True)
parser.add_argument("--output_mat", required=True)
parser.add_argument("--norm", required=True)
args = parser.parse_args()
var = sio.loadmat(args.mat)
in_data = torch.tensor(var[args.var_name])
with open(args.norm, "r") as f:
cfg = config.load_cfg(f)
input_mean = torch.tensor(cfg.input.mean)
input_std = torch.tensor(cfg.input.std)
target_mean = torch.tensor(cfg.target.mean)
target_std = torch.tensor(cfg.target.std)
in_data = (in_data - input_mean) / input_std
with open(args.weights, "rb") as f:
net = torch.load(f)
net.eval()
output = net(in_data) * target_std + target_mean
def get_cfg_from_file(self, fbase):
with open(fbase + "model_and_cfg/exp_cfg", "rt") as cfg_file:
return load_cfg(cfg_file)
def make_config(config_file):
with open(config_file, 'r') as f:
cfg = config.load_cfg(f)
return cfg.clone()
def read_yaml(filepath: str):
with open(filepath, "r") as f:
return load_cfg(f)