from transforms import *
import time
from datetime import datetime
# dataset load
transform = ComposeKeyPoints([
To3ChannelsIRKeyPoints(),
ResizeKeypoints(224),
RandomMirrorKeyPoints(),
RandomAffineKeyPoints((-60, 60)),
ToTensorKeyPoints(),
NormalizeKeyPoints((0.485, 0.456, 0.406), (0.229, 0.224, 0.225))
])
train_set = GaneratedHandsDataset(
"/disk1/ofirbartal/Projects/Dataset/GANeratedHands_Release/dataset_csv/train_dataset.csv",
transform)
train_loader = torch.utils.data.DataLoader(train_set,
batch_size=16,
shuffle=True)
val_set = GaneratedHandsDataset(
"/disk1/ofirbartal/Projects/Dataset/GANeratedHands_Release/dataset_csv/val_dataset.csv",
transform)
val_loader = torch.utils.data.DataLoader(val_set, batch_size=16, shuffle=True)
checkpoints_path = "/disk1/ofirbartal/Projects/LandmarksExtraction/Checkpoints/Autoencoder_checkpoints/{}/".format(
datetime.now().strftime("%d_%m_%y__%H_%M_%S"))
if not os.path.exists(checkpoints_path):
os.makedirs(checkpoints_path)
return np.array(dist)
# dataset load
transform = ComposeKeyPoints([
To3ChannelsGrayscaleKeyPoints(),
ResizeKeypoints(224),
RandomMirrorKeyPoints(),
RandomAffineKeyPoints((-60, 60)),
ToTensorKeyPoints(),
NormalizeKeyPoints((0.485, 0.456, 0.406), (0.229, 0.224, 0.225))
])
test_set = GaneratedHandsDataset(
"/disk1/ofirbartal/Projects/Dataset/GANeratedHands_Release/dataset_csv_small/test_dataset.csv",
transform)
test_loader = torch.utils.data.DataLoader(test_set,
batch_size=16,
shuffle=True)
pixel_error_threshhold_close = 5
pixel_error_threshhold_mid = 10
pixel_error_threshhold_far = 15
model = CoordRegressionNetwork(n_locations=21).cuda()
model = torch.nn.DataParallel(model, [0, 1], 0)
model.load_state_dict(
torch.load(
"/disk1/ofirbartal/Projects/LandmarksExtraction/Checkpoints/FCN_checkpoints/gray_26_11_19__14_36_37//epoch_80.pth"