def main(args):
print('===================================\n', )
print("Root directory: {}".format(args.name))
args.exp_dir = os.path.join(os.path.join(RES_DIR, args.name),
"evaluation_trinity_test")
if not os.path.isdir(args.exp_dir):
os.makedirs(args.exp_dir)
print("EXP PATH: {}".format(args.exp_dir))
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
args.device = device
assert args.checkpoint_seg is not None, "Need trained .pth!"
model_seg = load_models(
mode="segmentation",
device=device,
args=args,
)
assert args.checkpoint_disc is not None, "Need trained .pth!"
model_disc = load_models(
mode="single_discriminator",
device=device,
args=args,
)
transforms_shape_target = dual_transforms.Compose([
dual_transforms.CenterCrop((400, 400)),
dual_transforms.Scale(args.image_size[0]),
])
photo_transformer = PhotometricTransform(photometric_transform_config)
traindata = OpenEDSDataset_withLabels(
root=os.path.join(args.target_root, "train"),
image_size=args.image_size,
data_to_train="",
shape_transforms=transforms_shape_target,
photo_transforms=photo_transformer,
train_bool=False,
)
train_loader = torch.utils.data.DataLoader(
traindata,
batch_size=1,
shuffle=False,
num_workers=0,
pin_memory=True,
)
trainloader_iter = enumerate(train_loader)
args.total_iterations = traindata.__len__() // args.batch_size
model_seg.eval()
model_disc.eval()
confidence_map = []
confidence_mean_score = []
confidence_cnt = []
imagename = []
imageindex = []
dst_folder = os.path.join(args.exp_dir, "visualization")
if not os.path.isdir(dst_folder):
os.mkdir(dst_folder)
image_confidence = np.empty((traindata.__len__(), 224, 224, 2))
# pred_all = np.empty((traindata.__len__(),4,224,224))
# with open("dataloaders/eye/trinity_top_200.pkl", "rb") as f:
# top_200_lists = pickle.load(f)
with open("dataloaders/eye/top_1_adv.pkl", "rb") as f:
top_1_lists = pickle.load(f)
with open("dataloaders/eye/top_2_adv.pkl", "rb") as f:
top_2_lists = pickle.load(f)
for i_iter in range(args.total_iterations):
if i_iter % 1000 == 0:
print("Processing {} ..........".format(i_iter))
# if not (traindata.train_data_list[i_iter] in top_200_lists):
# continue
# if traindata.train_data_list[i_iter] in top_1_lists:
# continue
if traindata.train_data_list[i_iter] in top_2_lists:
continue
imageindex.append(i_iter)
imagename.append(traindata.train_data_list[i_iter])
_, batch = next(trainloader_iter)
images, labels = batch
images = Variable(images).to(args.device)
labels = Variable(labels.long()).to(args.device)
pred = model_seg(images)
pred_softmax = F.softmax(pred, dim=1)
D_out = model_disc(pred_softmax)
D_out = torch.sigmoid(D_out)
D_out = D_out[0, 0, :, :].detach().cpu().numpy()
pred = np.argmax(pred.detach().cpu().numpy(), axis=1)[0, :, :]
# fig = plt.figure()
# ax = fig.add_subplot(231)
# ax.imshow(images[0,0,:,:].detach().cpu().numpy(), cmap="gray")
# ax.set_xticks([])
# ax.set_yticks([])
#
# ax = fig.add_subplot(232)
# ax.imshow(labels[0, :, :].detach().cpu().numpy())
# ax.set_xticks([])
# ax.set_yticks([])
#
# ax = fig.add_subplot(233)
# ax.imshow(pred, cmap="gray")
# ax.set_xticks([])
# ax.set_yticks([])
#
# ax = fig.add_subplot(234)
# ax.imshow(D_out, cmap="gray")
# ax.set_xticks([])
# ax.set_yticks([])
D_out_mean = D_out.mean()
D_out_mean_map = (D_out > D_out_mean) * 1
# labels = labels[0,:,:].detach().cpu().numpy()
# semi_ignore_mask = (D_out < D_out_mean)
# # pseudo_gt = labels.copy()
# # pseudo_gt[semi_ignore_mask] = 4
# # pseudo_gt = pseudo_gt.astype(np.uint8)
# filename = traindata.train_data_list[i_iter].replace("/images/", "/masks/")
# filename = filename.replace("/train/", "/train_pseudo/")
# filename = filename.replace(".png", ".npy")
# np.save(filename, semi_ignore_mask)
# # print(D_out_mean_map.shape)
# ax = fig.add_subplot(235)
# ax.imshow(D_out_mean_map)
# ax.set_xticks([])
# ax.set_yticks([])
# plt.tight_layout()
# filename = ntpath.basename(traindata.train_data_list[i_iter])
# filename = os.path.join(dst_folder, filename)
# plt.savefig(filename)
# im_filename = traindata.train_data_list[i_iter].replace("/train/", "/train_pseudo/")
# os.system("cp %s %s" % (traindata.train_data_list[i_iter], im_filename))
confidence_mean_score.append(D_out_mean)
confidence_cnt.append(D_out_mean_map.sum())
### generate confidence map ###
# confidence_map.append(D_out_mean)
# imagename.append(ntpath.basename(traindata.train_data_list[i_iter]))
# image_confidence[i_iter,:,:,0] = images[0,0,:,:].detach().cpu().numpy()
# image_confidence[i_iter,:,:,1] = D_out
# pred_all[i_iter, ...] = pred_softmax[0,...].detach().cpu().numpy()
### generate confidence map ###
# with open("%s/confidence_map_top1_adv.pkl" % (args.exp_dir), "wb") as f:
# pickle.dump([imageindex, imagename, confidence_mean_score, confidence_cnt], f)
with open("%s/confidence_map_top2_adv.pkl" % (args.exp_dir), "wb") as f:
pickle.dump(
[imageindex, imagename, confidence_mean_score, confidence_cnt], f)
def main(args):
print('===================================\n', )
print("Root directory: {}".format(args.name))
args.exp_dir = os.path.join(RES_DIR, args.name)
if not os.path.isdir(args.exp_dir):
os.makedirs(args.exp_dir)
print("EXP PATH: {}".format(args.exp_dir))
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
args.device = device
logger = make_logger(filename="TrainVal.log", args=args)
if args.tensorboard:
writer = SummaryWriter(args.exp_dir)
else:
writer = None
print("===================================")
print("====== Loading Training Data ======")
print("===================================")
shape_transformer = dual_transforms.Compose([
dual_transforms.CenterCrop((400, 400)),
dual_transforms.Scale(args.image_size[0]),
])
photo_transformer = PhotometricTransform(photometric_transform_config)
source_data = UnityDataset(
root=args.source_root,
image_size=args.image_size,
shape_transforms=shape_transformer,
photo_transforms=photo_transformer,
train_bool=False,
)
args.tot_source = source_data.__len__()
args.total_iterations = args.num_epochs * args.tot_source // args.batch_size
args.iters_to_eval = args.epoch_to_eval * args.tot_source // args.batch_size
train_loader = torch.utils.data.DataLoader(
source_data,
batch_size=args.batch_size,
shuffle=True,
num_workers=args.workers,
pin_memory=True,
)
model_seg = load_models(
mode="segmentation",
device=device,
args=args,
)
optimizer_seg = optim.Adam(
model_seg.parameters(),
lr=args.lr_seg,
betas=(args.beta1, 0.999),
)
optimizer_seg.zero_grad()
seg_loss_target = torch.nn.CrossEntropyLoss().to(device)
trainloader_iter = enumerate(train_loader)
loss_seg_min = float("inf")
miou_max = float("-inf")
for i_iter in range(args.total_iterations):
loss_seg_value = 0
model_seg.train()
optimizer_seg.zero_grad()
adjust_learning_rate(
optimizer=optimizer_seg,
learning_rate=args.lr_seg,
i_iter=i_iter,
max_steps=args.total_iterations,
power=0.9,
)
try:
_, batch = next(trainloader_iter)
except StopIteration:
trainloader_iter = enumerate(train_loader)
_, batch = next(trainloader_iter)
images, labels = batch
images = Variable(images).to(args.device)
labels = Variable(labels.long()).to(args.device)
pred = model_seg(images)
loss_seg = seg_loss_target(pred, labels)
current_loss_seg = loss_seg.item()
loss_seg_value += current_loss_seg
loss_seg.backward()
optimizer_seg.step()
pred_img = pred.argmax(dim=1, keepdim=True)
flat_pred = pred_img.detach().cpu().numpy().flatten()
flat_gt = labels.detach().cpu().numpy().flatten()
miou, _ = compute_mean_iou(flat_pred=flat_pred, flat_label=flat_gt)
logger.info('iter = {0:8d}/{1:8d} '
'loss_seg = {2:.3f} '
'mIoU = {3:.3f} '.format(
i_iter,
args.total_iterations,
loss_seg_value,
miou,
))
if args.tensorboard and (writer != None):
writer.add_scalar('Train/Cross_Entropy', current_loss_seg, i_iter)
writer.add_scalar('Train/mIoU', miou, i_iter)
if i_iter % args.iters_to_eval == 0:
filename = os.path.join(
args.exp_dir, "Target_img_trainiter_{}.png".format(i_iter))
target_img = images.float()
gen_target_img = torchvision.utils.make_grid(target_img,
padding=2,
normalize=True)
torchvision.utils.save_image(gen_target_img, filename)
filename = os.path.join(
args.exp_dir, "Unity_pred_trainiter_{}.png".format(i_iter))
pred_img = pred_img.float()
gen_img = torchvision.utils.make_grid(pred_img,
padding=2,
normalize=True)
torchvision.utils.save_image(gen_img, filename)
is_better_ss = current_loss_seg < loss_seg_min
if is_better_ss:
loss_seg_min = current_loss_seg
torch.save(model_seg.state_dict(),
os.path.join(args.exp_dir, "model_train_best.pth"))
if miou > miou_max:
miou_max = miou
torch.save(model_seg.state_dict(),
os.path.join(args.exp_dir, "model_train_best_miou.pth"))
logger.info("==========================================")
logger.info("Training DONE!")
if args.tensorboard and (writer != None):
writer.close()
def main(args):
print('===================================\n', )
print("Root directory: {}".format(args.name))
args.exp_dir = os.path.join(os.path.join(RES_DIR, args.name),
"evaluation_trinity_test")
if not os.path.isdir(args.exp_dir):
os.makedirs(args.exp_dir)
print("EXP PATH: {}".format(args.exp_dir))
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
args.device = device
uncertainty_scores = np.load("uncertainty_scores.npy")
with (open(
"results/unity_to_trinity_UDA_semi_10/evaluation_trinity_test/confidence_map_UDA.pkl",
"rb")) as f:
adv_scores = pickle.load(f)
index = np.argsort(adv_scores[1])[::-1]
ascore = np.empty((2, 8916))
ascore[0] = index
ascore[1] = adv_scores[1][index]
assert args.checkpoint_seg is not None, "Need trained .pth!"
model_seg = load_models(
mode="segmentation",
device=device,
args=args,
)
# assert args.checkpoint_disc is not None, "Need trained .pth!"
# model_disc = load_models(
# mode="single_discriminator",
# device=device,
# args=args,
# )
class FeatureExtractor(torch.nn.Module):
def __init__(self, submodule, extracted_layers):
super(FeatureExtractor, self).__init__()
self.submodule = submodule
self.extracted_layers = extracted_layers
def forward(self, x):
for name, module in self.submodule._modules.items():
x = module(x)
print(name)
if name in self.extracted_layers:
return x['x5']
exact_list = ["pretrained_net"]
featExactor = FeatureExtractor(model_seg, exact_list)
# a = torch.randn(1, 3, 224, 224)
# a = Variable(a).to(args.device)
# x = myexactor(a)
# print(x)
transforms_shape_target = dual_transforms.Compose([
dual_transforms.CenterCrop((400, 400)),
dual_transforms.Scale(args.image_size[0]),
])
photo_transformer = PhotometricTransform(photometric_transform_config)
traindata = OpenEDSDataset_withLabels(
root=os.path.join(args.target_root, "train"),
image_size=args.image_size,
data_to_train="",
shape_transforms=transforms_shape_target,
photo_transforms=photo_transformer,
train_bool=False,
)
train_loader = torch.utils.data.DataLoader(
traindata,
batch_size=1,
shuffle=False,
num_workers=0,
pin_memory=True,
)
trainloader_iter = enumerate(train_loader)
args.total_iterations = traindata.__len__() // args.batch_size
model_seg.eval()
feature_maps = np.empty((traindata.__len__(), 25088))
count = 0
for i_iter in range(args.total_iterations):
if i_iter % 1000 == 0:
print("Processing {} ..........".format(i_iter))
# if uncertainty_scores[1][i_iter] == uncertainty_scores[1].max():
# print("uncertain", i_iter, traindata.train_data_list[i_iter], uncertainty_scores[1][i_iter], adv_scores[1][i_iter])
# count += 1
# if adv_scores[1][i_iter] == adv_scores[1].max():
# print("adv", i_iter, traindata.train_data_list[i_iter], uncertainty_scores[1][i_iter], adv_scores[1][i_iter])
# count += 1
#
# if count==2:
# break
# u_idx = np.where(uncertainty_scores[0]==i_iter)[0]
a_idx = np.where(ascore[0] == i_iter)[0]
_, batch = next(trainloader_iter)
images, labels = batch
images = Variable(images).to(args.device)
# labels = Variable(labels.long()).to(args.device)
feat = featExactor(images)
# feature_maps[i_iter] = feat.view(1,-1)[0].detach().cpu().numpy()
feature_maps[a_idx] = feat.view(1, -1)[0].detach().cpu().numpy()
# print("Saving feature maps .......................")
# np.save("feature_maps_UDA_original_order.npy", feature_maps)
# print("Saving feature maps .......................")
# np.save("feature_maps_UDA_ascore.npy", feature_maps)
from sklearn.metrics.pairwise import cosine_similarity
dist = cosine_similarity(feature_maps)
#
# A = np.matmul(feature_maps.transpose(), feature_maps)
# D = A.diagonal()
# distance_map = np.power(D, 0.5) * A * np.power(D, -0.5)
#
print("Saving distance maps .......................")
np.save("distance_maps_UDA_ascore.npy", dist)
def main(args):
print('===================================\n', )
print("Root directory: {}".format(args.name))
args.exp_dir = os.path.join(RES_DIR, args.name)
if not os.path.isdir(args.exp_dir):
os.makedirs(args.exp_dir)
print("EXP PATH: {}".format(args.exp_dir))
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
args.device = device
logger = make_logger(filename="TrainVal.log", args=args)
if args.tensorboard:
writer = SummaryWriter(args.exp_dir)
else:
writer = None
print("===================================")
print("====== Loading Training Data ======")
print("===================================")
shape_transformer = dual_transforms.Compose([
dual_transforms.CenterCrop((400, 400)),
dual_transforms.Scale(args.image_size[0]),
])
photo_transformer = PhotometricTransform(photometric_transform_config)
joint_data = JointDataset(
root_source=args.source_root,
image_size=args.image_size,
data_to_train="dataloaders/eye/trinity_train_200.pkl",
shape_transforms=shape_transformer,
photo_transforms=photo_transformer,
train_bool=False,
)
args.tot_data = joint_data.__len__()
args.total_iterations = args.num_epochs * args.tot_data // args.batch_size
args.iters_to_eval = args.epoch_to_eval * args.tot_data // args.batch_size
print("===================================")
print("========= Loading Val Data ========")
print("===================================")
val_target_data = OpenEDSDataset_withLabels(
root=os.path.join(args.target_root, "validation"),
image_size=args.image_size,
data_to_train="",
shape_transforms=shape_transformer,
photo_transforms=None,
train_bool=False,
)
train_loader = torch.utils.data.DataLoader(
joint_data,
batch_size=args.batch_size,
shuffle=True,
num_workers=args.workers,
pin_memory=True,
)
val_loader = torch.utils.data.DataLoader(
val_target_data,
batch_size=args.batch_size,
shuffle=True,
num_workers=args.workers,
pin_memory=True,
)
model_seg = load_models(
mode="segmentation",
device=device,
args=args,
)
optimizer_seg = optim.Adam(
model_seg.parameters(),
lr=args.lr_seg,
betas=(args.beta1, 0.999),
)
optimizer_seg.zero_grad()
# class_weight_target = 1.0 / train_target_data.get_class_probability().to(device)
seg_loss_target = torch.nn.CrossEntropyLoss().to(device)
trainloader_iter = enumerate(train_loader)
val_loss, val_miou = [], []
val_loss_f = float("inf")
val_miou_f = float("-inf")
loss_seg_min = float("inf")
for i_iter in range(args.total_iterations):
loss_seg_value = 0
model_seg.train()
optimizer_seg.zero_grad()
adjust_learning_rate(
optimizer=optimizer_seg,
learning_rate=args.lr_seg,
i_iter=i_iter,
max_steps=args.total_iterations,
power=0.9,
)
try:
_, batch = next(trainloader_iter)
except StopIteration:
trainloader_iter = enumerate(train_loader)
_, batch = next(trainloader_iter)
images, labels = batch
images = Variable(images).to(args.device)
labels = Variable(labels.long()).to(args.device)
pred = model_seg(images)
loss_seg = seg_loss_target(pred, labels)
current_loss_seg = loss_seg.item()
loss_seg_value += current_loss_seg
loss_seg.backward()
optimizer_seg.step()
logger.info('iter = {0:8d}/{1:8d} '
'loss_seg = {2:.3f} '.format(
i_iter,
args.total_iterations,
loss_seg_value,
))
current_epoch = i_iter * args.batch_size // args.tot_data
if i_iter % args.iters_to_eval == 0:
val_loss_f, val_miou_f = validate_baseline(
i_iter=i_iter,
val_loader=val_loader,
model=model_seg,
epoch=current_epoch,
logger=logger,
writer=writer,
val_loss=val_loss_f,
val_iou=val_miou_f,
args=args,
)
val_loss.append(val_loss_f)
val_loss_f = np.min(np.array(val_loss))
val_miou.append(val_miou_f)
val_miou_f = np.max(np.array(val_miou))
if args.tensorboard and (writer != None):
writer.add_scalar('Val/Cross_Entropy_Target', val_loss_f,
i_iter)
writer.add_scalar('Val/mIoU_Target', val_miou_f, i_iter)
is_better_ss = current_loss_seg < loss_seg_min
if is_better_ss:
loss_seg_min = current_loss_seg
torch.save(model_seg.state_dict(),
os.path.join(args.exp_dir, "model_train_best.pth"))
logger.info("==========================================")
logger.info("Training DONE!")
if args.tensorboard and (writer != None):
writer.close()
with open("%s/train_performance.pkl" % args.exp_dir, "wb") as f:
pickle.dump([val_loss, val_miou], f)
logger.info("==========================================")
logger.info("Evaluating on test data ...")
testdata = OpenEDSDataset_withLabels(
root=os.path.join(args.target_root, "test"),
image_size=args.image_size,
data_to_train="",
shape_transforms=shape_transformer,
photo_transforms=None,
train_bool=False,
)
test_loader = torch.utils.data.DataLoader(
testdata,
batch_size=1,
shuffle=False,
num_workers=0,
pin_memory=True,
)
pm = run_testing(
dataset=testdata,
test_loader=test_loader,
model=model_seg,
args=args,
)
logger.info('Global Mean Accuracy: {:.3f}'.format(np.array(pm.GA).mean()))
logger.info('Mean IOU: {:.3f}'.format(np.array(pm.IOU).mean()))
logger.info('Mean Recall: {:.3f}'.format(np.array(pm.Recall).mean()))
logger.info('Mean Precision: {:.3f}'.format(np.array(pm.Precision).mean()))
logger.info('Mean F1: {:.3f}'.format(np.array(pm.F1).mean()))
IOU_ALL = np.array(pm.Iou_all)
logger.info(
"Back: {:.4f}, Sclera: {:.4f}, Iris: {:.4f}, Pupil: {:.4f}".format(
IOU_ALL[:, 0].mean(),
IOU_ALL[:, 1].mean(),
IOU_ALL[:, 2].mean(),
IOU_ALL[:, 3].mean(),
))
def main(args):
print('===================================\n', )
print("Root directory: {}".format(args.name))
args.exp_dir = os.path.join(RES_DIR, args.name)
if not os.path.isdir(args.exp_dir):
os.makedirs(args.exp_dir)
print("EXP PATH: {}".format(args.exp_dir))
assert args.trinity_data_train_with_labels != "", "Indicate trained data in Trinity!"
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
args.device = device
logger = make_logger(filename="TrainVal.log", args=args)
if args.tensorboard:
writer = SummaryWriter(args.exp_dir)
else:
writer = None
print("===================================")
print("====== Loading Training Data ======")
print("===================================")
transforms_shape_source = dual_transforms.Compose([
dual_transforms.CenterCrop((400, 400)),
dual_transforms.Scale(args.image_size[0]),
])
transforms_shape_target = dual_transforms.Compose([
dual_transforms.CenterCrop((400, 400)),
dual_transforms.Scale(args.image_size[0]),
])
photo_transformer = PhotometricTransform(photometric_transform_config)
source_data = UnityDataset(
root=args.source_root,
image_size=args.image_size,
shape_transforms=transforms_shape_source,
photo_transforms=photo_transformer,
train_bool=False,
)
target_data = OpenEDSDataset_withLabels(
root=os.path.join(args.target_root, "train"),
image_size=args.image_size,
data_to_train=args.trinity_data_train_with_labels,
shape_transforms=transforms_shape_target,
photo_transforms=photo_transformer,
train_bool=False,
)
args.tot_source = source_data.__len__()
args.total_iterations = args.num_epochs * source_data.__len__(
) // args.batch_size
args.iters_to_eval = args.epoch_to_eval * source_data.__len__(
) // args.batch_size
args.iter_source_to_eval = args.epoch_to_eval_source * source_data.__len__(
) // args.batch_size
args.iter_semi_start = args.epoch_semi_start * source_data.__len__(
) // args.batch_size
print("===================================")
print("========= Loading Val Data ========")
print("===================================")
val_target_data = OpenEDSDataset_withLabels(
root=os.path.join(args.target_root, "validation"),
image_size=args.image_size,
data_to_train="",
shape_transforms=transforms_shape_target,
photo_transforms=None,
train_bool=False,
)
# class_weight_source = 1.0 / source_data.get_class_probability().to(device)
source_loader = torch.utils.data.DataLoader(
source_data,
batch_size=args.batch_size,
shuffle=True,
num_workers=args.workers,
pin_memory=True,
)
target_loader = torch.utils.data.DataLoader(
target_data,
batch_size=args.batch_size,
shuffle=True,
num_workers=args.workers,
pin_memory=True,
)
val_loader = torch.utils.data.DataLoader(
val_target_data,
batch_size=args.batch_size,
shuffle=True,
num_workers=args.workers,
pin_memory=True,
)
model_seg = load_models(
mode="segmentation",
device=device,
args=args,
)
model_disc = load_models(
mode="single_discriminator",
device=device,
args=args,
)
optimizer_seg = optim.Adam(
model_seg.parameters(),
lr=args.lr_seg,
betas=(args.beta1, 0.999),
)
optimizer_seg.zero_grad()
optimizer_disc = torch.optim.Adam(
model_disc.parameters(),
lr=args.lr_disc,
betas=(args.beta1, 0.999),
)
optimizer_disc.zero_grad()
seg_loss_source = torch.nn.CrossEntropyLoss().to(device)
gan_loss = torch.nn.BCEWithLogitsLoss().to(device)
semi_loss = torch.nn.CrossEntropyLoss(ignore_index=-1)
history_true_mask = ImagePool(args.pool_size)
history_fake_mask = ImagePool(args.pool_size)
trainloader_iter = enumerate(source_loader)
targetloader_iter = enumerate(target_loader)
val_loss, val_miou = run_training_SSDA(
trainloader_source=source_loader,
trainloader_target=target_loader,
trainloader_iter=trainloader_iter,
targetloader_iter=targetloader_iter,
val_loader=val_loader,
model_seg=model_seg,
model_disc=model_disc,
gan_loss=gan_loss,
seg_loss=seg_loss_source,
semi_loss_criterion=semi_loss,
optimizer_seg=optimizer_seg,
optimizer_disc=optimizer_disc,
history_pool_true=history_true_mask,
history_pool_fake=history_fake_mask,
logger=logger,
writer=writer,
args=args,
)
with open("%s/train_performance.pkl" % args.exp_dir, "wb") as f:
pickle.dump([val_loss, val_miou], f)
logger.info("==========================================")
logger.info("Evaluating on test data ...")
testdata = OpenEDSDataset_withLabels(
root=os.path.join(args.target_root, "test"),
image_size=args.image_size,
data_to_train="",
shape_transforms=transforms_shape_target,
photo_transforms=None,
train_bool=False,
)
test_loader = torch.utils.data.DataLoader(
testdata,
batch_size=1,
shuffle=False,
num_workers=0,
pin_memory=True,
)
pm = run_testing(
dataset=testdata,
test_loader=test_loader,
model=model_seg,
args=args,
)
logger.info('Global Mean Accuracy: {:.3f}'.format(np.array(pm.GA).mean()))
logger.info('Mean IOU: {:.3f}'.format(np.array(pm.IOU).mean()))
logger.info('Mean Recall: {:.3f}'.format(np.array(pm.Recall).mean()))
logger.info('Mean Precision: {:.3f}'.format(np.array(pm.Precision).mean()))
logger.info('Mean F1: {:.3f}'.format(np.array(pm.F1).mean()))
IOU_ALL = np.array(pm.Iou_all)
logger.info(
"Back: {:.4f}, Sclera: {:.4f}, Iris: {:.4f}, Pupil: {:.4f}".format(
IOU_ALL[:, 0].mean(),
IOU_ALL[:, 1].mean(),
IOU_ALL[:, 2].mean(),
IOU_ALL[:, 3].mean(),
))