optimizer = torch.optim.Adam(model.parameters(), lr=0.001)
train_batch_logger = Logger(os.path.join('./Results', 'train_batch.log'), ['epoch', 'batch', 'loss', 'acc'])
test_batch_logger = Logger(os.path.join('./Results', 'test_batch.log'), ['batch', 'loss', 'acc'])
acc_logger = Logger(os.path.join('./Results', 'acc.log'), ['epoch', 'acc'])
#shutil.rmtree(osp.join('..', 'data/Traingraph/processed'))
#shutil.rmtree(osp.join('..', 'data/Testgraph/processed'))
for epoch in range(1, 180):
train_path = osp.join('..', 'data/Traingraph')
test_path = osp.join('..', 'data/Testgraph')
train_data_aug = T.Compose([T.Cartesian(cat=False), T.RandomFlip(axis=0, p=0.3), T.RandomScale([0.95,0.999]) ])
test_data_aug = T.Compose([T.Cartesian(cat=False), T.RandomFlip(axis=0, p=0.3), T.RandomScale([0.95,0.999])])
train_dataset = MyOwnDataset(train_path, transform=train_data_aug) #### transform=T.Cartesian()
test_dataset = MyOwnDataset(test_path, transform=test_data_aug)
train_loader = DataLoader(train_dataset, batch_size=64, shuffle=True)
test_loader = DataLoader(test_dataset, batch_size=64)
train(epoch, train_batch_logger, train_loader)
test_acc = test(test_batch_logger, test_loader)
print('Epoch: {:02d}, Test: {:.4f}'.format(epoch, test_acc))
torch.save(model, './runs_model/model.pkl')
# ## Data loading
# Let's get the dataset
# In[2]:
import torch
from torch_geometric.datasets import ModelNet
import torch_geometric.transforms as T
import time
from tqdm import tqdm_notebook
pre_transform = T.NormalizeScale()
transform = T.Compose([
T.SamplePoints(1024),
T.RandomRotate(30),
T.RandomScale((0.5, 2)),
])
name = '40'
train_ds = ModelNet(root='./',
train=True,
name=name,
pre_transform=pre_transform,
transform=transform)
test_ds = ModelNet(root='./',
train=True,
name=name,
pre_transform=pre_transform,
transform=T.SamplePoints(1024 * 4))
args = parser.parse_args()
torch.multiprocessing.set_sharing_strategy('file_system')
torch.backends.cudnn.benchmark = False
logging.basicConfig(filename=Config.graph_train_log_path,
level=logging.DEBUG)
device = torch.device("cuda:" +
args.cuda if torch.cuda.is_available() else "cpu")
model = Net()
model = model.to(device)
model.load_state_dict(
torch.load(os.path.join(Config.model_dir, args.model_name)))
test_data_aug = T.Compose(
[T.Cartesian(cat=False),
T.RandomScale([0.99999, 1])])
# test_data_aug = T.Compose([T.Cartesian(cat=False)])
test_dataset = EV_Gait_3DGraph_Dataset(Config.graph_test_dir,
transform=test_data_aug)
test_loader = DataLoader(test_dataset,
batch_size=16,
num_workers=2,
pin_memory=True)
# test
model.eval()
correct = 0
total = 0
for index, data in enumerate(tqdm(test_loader)):
data = data.to(device)
test_batch_logger = Logger(os.path.join('./Results', 'test_batch.log'),
['batch', 'loss', 'acc'])
acc_logger = Logger(os.path.join('./Results', 'acc.log'), ['epoch', 'acc'])
#shutil.rmtree(osp.join('..', 'data/Traingraph/processed'))
#shutil.rmtree(osp.join('..', 'data/Testgraph/processed'))
for epoch in range(1, 180):
train_path = osp.join('..', 'data/Traingraph')
test_path = osp.join('..', 'data/Testgraph')
#train_data_aug = T.Compose([T.Cartesian(cat=False), T.RandomFlip(axis=0)])
train_data_aug = T.Compose([
T.Cartesian(cat=False),
T.RandomFlip(axis=0, p=0.3),
T.RandomScale([0.96, 0.999])
])
test_data_aug = T.Compose([
T.Cartesian(cat=False),
T.RandomFlip(axis=0, p=0.3),
T.RandomScale([0.96, 0.999])
])
train_dataset = MyOwnDataset(
train_path, transform=train_data_aug) #### transform=T.Cartesian()
test_dataset = MyOwnDataset(test_path, transform=test_data_aug)
train_loader = DataLoader(train_dataset, batch_size=3, shuffle=True)
test_loader = DataLoader(test_dataset, batch_size=3)
train(epoch, train_batch_logger, train_loader)
train_batch_logger = Logger(os.path.join('./Results', 'train_batch.log'),
['epoch', 'batch', 'loss', 'acc'])
test_batch_logger = Logger(os.path.join('./Results', 'test_batch.log'),
['batch', 'loss', 'acc'])
#shutil.rmtree(osp.join('..', 'data/Traingraph/processed'))
#shutil.rmtree(osp.join('..', 'data/Testgraph/processed'))
for epoch in range(1, 150):
train_path = osp.join('..', 'data/Traingraph')
test_path = osp.join('..', 'data/Testgraph')
train_data_aug = T.Compose([
T.Cartesian(cat=False),
T.RandomFlip(axis=0, p=0.3),
T.RandomScale([0.95, 0.999]),
T.RandomFlip(axis=1, p=0.2)
])
test_data_aug = T.Compose([
T.Cartesian(cat=False),
T.RandomFlip(axis=0, p=0.3),
T.RandomScale([0.95, 0.999]),
T.RandomFlip(axis=1, p=0.2)
])
train_dataset = MyOwnDataset(
train_path, transform=train_data_aug) #### transform=T.Cartesian()
test_dataset = MyOwnDataset(test_path, transform=test_data_aug)
train_loader = DataLoader(train_dataset, batch_size=3, shuffle=True)
test_loader = DataLoader(test_dataset, batch_size=3)
os.makedirs(model_dir)
logging.basicConfig(filename=os.path.join(log_dir, 'n_mnist.log'),
level=logging.DEBUG)
model_file = 'n_mnist.pkl'
device = torch.device("cuda:" +
args.cuda if torch.cuda.is_available() else "cpu")
model = model.to(device)
optimizer = torch.optim.Adam(model.parameters(), lr=0.001)
pre_transform = T.Compose([T.Cartesian(cat=False)])
# train_data_aug = T.Compose([T.RandomScale([0.95, 1])])
train_data_aug = T.Compose([
T.RandomScale([0.95, 1]),
T.RandomRotate((0, 10), axis=0),
T.RandomFlip(axis=0, p=0.5)
])
train_dataset = Graph_2D_Memory_Dataset(train_dir,
transform=train_data_aug,
pre_transform=pre_transform)
train_loader = DataLoader(train_dataset,
batch_size=args.batch_size,
shuffle=True,
num_workers=args.workers)
# train
print("train")
for epoch in range(1, args.epoch):
parser = argparse.ArgumentParser()
parser.add_argument("--cuda", default="0", help="The GPU ID")
parser.add_argument("--epoch", default=150, type=int, help="The GPU ID")
parser.add_argument("--batch_size", default=16, type=int, help="batch size")
args = parser.parse_args()
torch.multiprocessing.set_sharing_strategy('file_system')
torch.backends.cudnn.benchmark = False
logging.basicConfig(filename=Config.graph_train_log_path, level=logging.DEBUG)
device = torch.device("cuda:" + args.cuda if torch.cuda.is_available() else "cpu")
model = Net()
model = model.to(device)
optimizer = torch.optim.Adam(model.parameters(), lr=0.001)
# train_data_aug = T.Compose([T.Cartesian(cat=False), T.RandomScale([0.96, 0.999]), T.RandomTranslate(0.01)])
train_data_aug = T.Compose([T.Cartesian(cat=False), T.RandomScale([0.96, 0.999])])
train_dataset = EV_Gait_3DGraph_Dataset(
Config.graph_train_dir, transform=train_data_aug
)
train_loader = DataLoader(train_dataset, batch_size=args.batch_size, shuffle=True, num_workers=2, pin_memory=True)
test_data_aug = T.Compose([T.Cartesian(cat=False), T.RandomScale([0.999, 1])])
# test_data_aug = T.Compose([T.Cartesian(cat=False)])
test_dataset = EV_Gait_3DGraph_Dataset(
Config.graph_test_dir, transform=test_data_aug
)
test_loader = DataLoader(test_dataset, batch_size=args.batch_size, num_workers=2, pin_memory=True)
# train