batch_size = 32
num_workers = 12
mean = [-17.398721187929123, -10.020421713800838, -12.10841437771272]
std = [6.290316422115964, 5.776936185931195, 5.795418280085563]
max_value = 1.0
transforms = A.Compose(
[A.Normalize(mean=mean, std=std, max_pixel_value=max_value),
ToTensorV2()])
_, data_loader, _ = get_train_val_loaders(
train_ds,
val_ds,
train_transforms=transforms,
val_transforms=transforms,
batch_size=batch_size,
num_workers=num_workers,
val_batch_size=batch_size,
pin_memory=True,
)
prepare_batch = inference_prepare_batch_f32
# Image denormalization function to plot predictions with images
img_denormalize = partial(denormalize, mean=mean, std=std)
#################### Model ####################
model = FPN(encoder_name='se_resnext50_32x4d', classes=2, encoder_weights=None)
run_uuid = "5230c20f609646cb9870a211036ea5cb"
weights_filename = "best_model_67_val_miou_bg=0.7574240313552584.pth"
A.Flip(),
]),
A.Normalize(mean=mean, std=std, max_pixel_value=max_value),
ToTensorV2()
])
val_transforms = A.Compose(
[A.Normalize(mean=mean, std=std, max_pixel_value=max_value),
ToTensorV2()])
train_loader, val_loader, train_eval_loader = get_train_val_loaders(
train_ds,
val_ds,
train_transforms=train_transforms,
val_transforms=val_transforms,
batch_size=batch_size,
num_workers=num_workers,
val_batch_size=val_batch_size,
pin_memory=True,
train_sampler=train_sampler,
limit_train_num_samples=100 if debug else None,
limit_val_num_samples=100 if debug else None)
accumulation_steps = 2
prepare_batch = prepare_batch_fp32
# Image denormalization function to plot predictions with images
def img_denormalize(nimg):
img = denormalize(nimg, mean=mean, std=std)
return img[(0, 1, 2), :, :]
val_transforms = A.Compose(
[
A.PadIfNeeded(val_img_size, val_img_size, border_mode=cv2.BORDER_CONSTANT),
A.Normalize(mean=mean, std=std),
ignore_mask_boundaries,
ToTensor(),
]
)
train_loader, val_loader, train_eval_loader = get_train_val_loaders(
root_path=data_path,
train_transforms=train_transforms,
val_transforms=val_transforms,
batch_size=batch_size,
num_workers=num_workers,
val_batch_size=val_batch_size,
limit_train_num_samples=100 if debug else None,
limit_val_num_samples=100 if debug else None,
)
prepare_batch = prepare_batch_fp32
# Image denormalization function to plot predictions with images
img_denormalize = partial(denormalize, mean=(0.485, 0.456, 0.406), std=(0.229, 0.224, 0.225))
# ##############################
# Setup models
# ##############################
num_classes = 21
])
val_transforms = A.Compose([
A.PadIfNeeded(val_img_size, val_img_size, border_mode=cv2.BORDER_CONSTANT),
A.Normalize(mean=mean, std=std),
ignore_mask_boundaries,
ToTensor(),
])
train_loader, val_loader, train_eval_loader = get_train_val_loaders(
root_path=data_path,
train_transforms=train_transforms,
val_transforms=val_transforms,
batch_size=batch_size,
num_workers=num_workers,
val_batch_size=val_batch_size,
with_sbd=sbd_data_path,
train_sampler='distributed',
val_sampler='distributed',
limit_train_num_samples=100 if debug else None,
limit_val_num_samples=100 if debug else None,
random_seed=seed)
prepare_batch = prepare_batch_fp32
# Image denormalization function to plot predictions with images
img_denormalize = partial(denormalize,
mean=(0.485, 0.456, 0.406),
std=(0.229, 0.224, 0.225))
# ##############################
ToTensor(),
])
val_transforms = A.Compose([
# https://github.com/facebookresearch/FixRes/blob/b27575208a7c48a3a6e0fa9efb57baa4021d1305/imnet_resnet50_scratch/transforms.py#L76
A.Resize(int((256 / 224) * val_crop_size), int(
(256 / 224) * val_crop_size)),
A.CenterCrop(val_crop_size, val_crop_size),
A.Normalize(mean=mean, std=std),
ToTensor(),
])
train_loader, val_loader, train_eval_loader = get_train_val_loaders(
data_path,
train_transforms=train_transforms,
val_transforms=val_transforms,
batch_size=batch_size,
num_workers=num_workers,
val_batch_size=batch_size,
)
# Image denormalization function to plot predictions with images
img_denormalize = partial(denormalize, mean=mean, std=std)
# ##############################
# Setup Model
# ##############################
model = resnet50(pretrained=False)
# ##############################
# Setup Solver
val_transforms = A.Compose([
# https://github.com/facebookresearch/FixRes/blob/b27575208a7c48a3a6e0fa9efb57baa4021d1305/imnet_resnet50_scratch/transforms.py#L76
A.Resize(int((256 / 224) * val_crop_size), int(
(256 / 224) * val_crop_size)),
A.CenterCrop(val_crop_size, val_crop_size),
A.Normalize(mean=mean, std=std),
ToTensor(),
])
train_loader, val_loader, train_eval_loader = get_train_val_loaders(
data_path,
train_transforms=train_transforms,
val_transforms=val_transforms,
batch_size=batch_size,
num_workers=num_workers,
val_batch_size=batch_size,
pin_memory=True,
train_sampler="distributed",
val_sampler="distributed",
)
# Image denormalization function to plot predictions with images
img_denormalize = partial(denormalize, mean=mean, std=std)
# ##############################
# Setup Model
# ##############################
model = resnet50(pretrained=False)
CenterCrop(img_size, img_size),
Normalize(mean=(0.485, 0.456, 0.406), std=(0.229, 0.224, 0.225)),
ignore_mask_boundaries,
ToTensor(),
])
val_transform_fn = lambda dp: val_transforms(**dp)
batch_size = 8
num_workers = 10
non_blocking = True
train_loader, val_loader, train_eval_loader = get_train_val_loaders(
root_path=data_path,
train_transforms=train_transform_fn,
val_transforms=val_transform_fn,
batch_size=batch_size,
num_workers=num_workers,
val_batch_size=batch_size * 2,
limit_train_num_samples=250 if debug else None,
limit_val_num_samples=250 if debug else None,
random_seed=seed)
prepare_batch = prepare_batch_fp32
val_interval = 5
num_classes = 21
model = DeepLabV3(build_resnet18_backbone, num_classes=num_classes)
criterion = nn.CrossEntropyLoss()
lr = 0.007 / 4.0 * batch_size
data_dir = Path('/media/wwymak/Storage/spacenet/AOI_3_Paris_Train')
image_dir = data_dir / 'RGB-PanSharpen'
mask_dir = data_dir / 'masks'
summary_data_filepath = data_dir / 'summaryData' / 'AOI_3_Paris_Train_Building_Solutions.csv'
mean = (0.485, 0.456, 0.406)
std = (0.229, 0.224, 0.225)
train_transforms = get_train_augmentation()
val_transforms = get_validation_augmentation()
train_loader, val_loader, train_eval_loader = get_train_val_loaders(summary_data_filepath=summary_data_filepath,
train_transforms=get_train_augmentation(),
val_transforms=get_validation_augmentation(),
batch_size=batch_size, num_workers=num_workers,
limit_train_num_samples=100 if debug else None,
limit_val_num_samples=100 if debug else None)
prepare_batch = prepare_batch_fp32
# Image denormalization function to plot predictions with images
img_denormalize = partial(denormalize, mean=(0.485, 0.456, 0.406), std=(0.229, 0.224, 0.225))
# ##############################
# Setup models
# ##############################
num_classes = 1
# model = deeplabv3_resnet101(num_classes=num_classes)
model, preprocessing_function = unet_resnet('efficientnet-b0')
batch_size = 4
num_workers = 12
if __name__ == "__main__":
print('here')
model, preprocessing_fn = unet_resnet(ENCODER)
runner = SupervisedRunner(device="cuda:0", input_key="image", input_target_key="mask", model=model)
summary_data_df = pd.read_csv(summary_data)
ground_truth_data_df = pd.read_csv(ground_truth_data)
train_loader, val_loader, test_loader = get_train_val_loaders(
summary_data_filepath=summary_data,
train_transforms=get_train_augmentation(is_display=False, image_size=512, normalisation_fn=None),
val_transforms=get_validation_augmentation(is_display=False, image_size=512, normalisation_fn=None),
batch_size=batch_size,
num_workers=num_workers, drop_empty_images=False
)
predictions_all = []
image_filepaths_all = []
image_ids_all = []
for loader in [train_loader, val_loader, test_loader]:
predictions = np.vstack(list(map(
lambda x: x["logits"].cpu().numpy(),
runner.predict_loader(loader=loader, resume=f"{logdir}/checkpoints/best.pth")
)))
image_filepaths = loader.dataset.image_filepath_list
predictions_all.append(predictions)
