Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Replace deprecated np.int with np.int_ and np.float with np.float64 #1659

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Replace deprecated np.int with np.int_ and np.float with np.float64 #1659

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
28 changes: 14 additions & 14 deletions pcdet/datasets/augmentor/database_sampler.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -178,24 +178,24 @@ def copy_paste_to_image_kitti(self, data_dict, crop_feat, gt_number, point_idxes
paste_order = boxes3d[:,0].argsort()
paste_order = paste_order[::-1]
else:
paste_order = np.arange(len(boxes3d),dtype=np.int)
paste_order = np.arange(len(boxes3d),dtype=np.int_)

if 'reverse' in kitti_img_aug_type:
paste_order = paste_order[::-1]

paste_mask = -255 * np.ones(image.shape[:2], dtype=np.int)
fg_mask = np.zeros(image.shape[:2], dtype=np.int)
overlap_mask = np.zeros(image.shape[:2], dtype=np.int)
depth_mask = np.zeros((*image.shape[:2], 2), dtype=np.float)
paste_mask = -255 * np.ones(image.shape[:2], dtype=np.int_)
fg_mask = np.zeros(image.shape[:2], dtype=np.int_)
overlap_mask = np.zeros(image.shape[:2], dtype=np.int_)
depth_mask = np.zeros((*image.shape[:2], 2), dtype=np.float64)
points_2d, depth_2d = data_dict['calib'].lidar_to_img(data_dict['points'][:,:3])
points_2d[:,0] = np.clip(points_2d[:,0], a_min=0, a_max=image.shape[1]-1)
points_2d[:,1] = np.clip(points_2d[:,1], a_min=0, a_max=image.shape[0]-1)
points_2d = points_2d.astype(np.int)
points_2d = points_2d.astype(np.int_)
for _order in paste_order:
_box2d = boxes2d[_order]
image[_box2d[1]:_box2d[3],_box2d[0]:_box2d[2]] = crop_feat[_order]
overlap_mask[_box2d[1]:_box2d[3],_box2d[0]:_box2d[2]] += \
(paste_mask[_box2d[1]:_box2d[3],_box2d[0]:_box2d[2]] > 0).astype(np.int)
(paste_mask[_box2d[1]:_box2d[3],_box2d[0]:_box2d[2]] > 0).astype(np.int_)
paste_mask[_box2d[1]:_box2d[3],_box2d[0]:_box2d[2]] = _order

if 'cover' in kitti_img_aug_use_type:
Expand Down Expand Up @@ -255,22 +255,22 @@ def collect_image_crops_kitti(self, info, data_dict, obj_points, sampled_gt_boxe
sampled_gt_boxes[idx] = box3d_lidar[0]
sampled_gt_boxes2d[idx] = box2d[0]

obj_idx = idx * np.ones(len(obj_points), dtype=np.int)
obj_idx = idx * np.ones(len(obj_points), dtype=np.int_)

# copy crops from images
img_path = self.root_path / f'training/image_2/{info["image_idx"]}.png'
raw_image = io.imread(img_path)
raw_image = raw_image.astype(np.float32)
raw_center = info['bbox'].reshape(2,2).mean(0)
new_box = sampled_gt_boxes2d[idx].astype(np.int)
new_box = sampled_gt_boxes2d[idx].astype(np.int_)
new_shape = np.array([new_box[2]-new_box[0], new_box[3]-new_box[1]])
raw_box = np.concatenate([raw_center-new_shape/2, raw_center+new_shape/2]).astype(np.int)
raw_box = np.concatenate([raw_center-new_shape/2, raw_center+new_shape/2]).astype(np.int_)
raw_box[0::2] = np.clip(raw_box[0::2], a_min=0, a_max=raw_image.shape[1])
raw_box[1::2] = np.clip(raw_box[1::2], a_min=0, a_max=raw_image.shape[0])
if (raw_box[2]-raw_box[0])!=new_shape[0] or (raw_box[3]-raw_box[1])!=new_shape[1]:
new_center = new_box.reshape(2,2).mean(0)
new_shape = np.array([raw_box[2]-raw_box[0], raw_box[3]-raw_box[1]])
new_box = np.concatenate([new_center-new_shape/2, new_center+new_shape/2]).astype(np.int)
new_box = np.concatenate([new_center-new_shape/2, new_center+new_shape/2]).astype(np.int_)

img_crop2d = raw_image[raw_box[1]:raw_box[3],raw_box[0]:raw_box[2]] / 255

Expand Down Expand Up @@ -320,8 +320,8 @@ def initilize_image_aug_dict(self, data_dict, gt_boxes_mask):
pass
elif self.img_aug_type == 'kitti':
obj_index_list, crop_boxes2d = [], []
gt_number = gt_boxes_mask.sum().astype(np.int)
gt_boxes2d = data_dict['gt_boxes2d'][gt_boxes_mask].astype(np.int)
gt_number = gt_boxes_mask.sum().astype(np.int_)
gt_boxes2d = data_dict['gt_boxes2d'][gt_boxes_mask].astype(np.int_)
gt_crops2d = [data_dict['images'][_x[1]:_x[3],_x[0]:_x[2]] for _x in gt_boxes2d]

img_aug_gt_dict = {
Expand Down Expand Up @@ -351,7 +351,7 @@ def collect_image_crops(self, img_aug_gt_dict, info, data_dict, obj_points, samp
def copy_paste_to_image(self, img_aug_gt_dict, data_dict, points):
if self.img_aug_type == 'kitti':
obj_points_idx = np.concatenate(img_aug_gt_dict['obj_index_list'], axis=0)
point_idxes = -1 * np.ones(len(points), dtype=np.int)
point_idxes = -1 * np.ones(len(points), dtype=np.int_)
point_idxes[:obj_points_idx.shape[0]] = obj_points_idx

data_dict['gt_boxes2d'] = np.concatenate([img_aug_gt_dict['gt_boxes2d'], np.array(img_aug_gt_dict['crop_boxes2d'])], axis=0)
Expand Down
6 changes: 3 additions & 3 deletions pcdet/models/backbones_2d/base_bev_backbone.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -57,7 +57,7 @@ def __init__(self, model_cfg, input_channels):
nn.ReLU()
))
else:
stride = np.round(1 / stride).astype(np.int)
stride = np.round(1 / stride).astype(np.int_)
self.deblocks.append(nn.Sequential(
nn.Conv2d(
num_filters[idx], num_upsample_filters[idx],
Expand Down Expand Up @@ -158,7 +158,7 @@ def __init__(self, model_cfg, **kwargs):
nn.ReLU()
))
else:
stride = np.round(1 / stride).astype(np.int)
stride = np.round(1 / stride).astype(np.int_)
self.deblocks.append(nn.Sequential(
nn.Conv2d(
num_filters[idx], num_upsample_filters[idx],
Expand Down Expand Up @@ -296,7 +296,7 @@ def __init__(self, model_cfg, input_channels):
nn.ReLU()
))
else:
stride = np.round(1 / stride).astype(np.int)
stride = np.round(1 / stride).astype(np.int_)
self.deblocks.append(nn.Sequential(
nn.Conv2d(
num_filters[idx], num_upsample_filters[idx],
Expand Down