Open xiatian815 opened 11 months ago
$ python main_HAR.py Traceback (most recent call last): File "main_HAR.py", line 166, in train = Train(args) File "main_HAR.py", line 22, in init self.train, self.valid, self.test = data_generator('E:\FC_STGNN\data\HAR', args=args) File "E:\FC_STGNN\data_loader_HAR.py", line 49, in data_generator train_dataset = Load_Dataset(train_dataset, args) File "E:\FC_STGNN\data_loader_HAR.py", line 32, in init self.x_data = self.x_data.reshape(shape[0],shape[1],args.time_denpen_len, args.window_size) AttributeError: 'Namespace' object has no attribute 'window_size' (gpu-torch)
Sorry for the late reply. This hyper-parameter 'window_size' should be 'patch_size'. Thanks for pointing out this problem.
We have updated and revised this problem.
$ python main_HAR.py Traceback (most recent call last): File "main_HAR.py", line 166, in
train = Train(args)
File "main_HAR.py", line 22, in init
self.train, self.valid, self.test = data_generator('E:\FC_STGNN\data\HAR', args=args)
File "E:\FC_STGNN\data_loader_HAR.py", line 49, in data_generator
train_dataset = Load_Dataset(train_dataset, args)
File "E:\FC_STGNN\data_loader_HAR.py", line 32, in init
self.x_data = self.x_data.reshape(shape[0],shape[1],args.time_denpen_len, args.window_size)
AttributeError: 'Namespace' object has no attribute 'window_size'
(gpu-torch)