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

Fix mask test with mismatched device #511

Merged
merged 3 commits into from
May 26, 2022
Merged

Fix mask test with mismatched device #511

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
237cdf5
align mask output to cpu device
RunningLeon May 23, 2022
aeacf62
align ncnn ssd output to torch.Tensor type
RunningLeon May 23, 2022
5f486cf
--amend
RunningLeon May 23, 2022
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
53 changes: 24 additions & 29 deletions mmdeploy/codebase/mmdet/deploy/object_detection_model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -111,37 +111,40 @@ def __clear_outputs(
return outputs

@staticmethod
def postprocessing_masks(det_bboxes: np.ndarray,
det_masks: np.ndarray,
def postprocessing_masks(det_bboxes: Union[np.ndarray, torch.Tensor],
det_masks: Union[np.ndarray, torch.Tensor],
img_w: int,
img_h: int,
device: str = 'cpu',
mask_thr_binary: float = 0.5) -> np.ndarray:
device: str = 'cpu') -> torch.Tensor:
"""Additional processing of masks. Resizes masks from [num_det, 28, 28]
to [num_det, img_w, img_h]. Analog of the 'mmdeploy.codebase.mmdet.
models.roi_heads.fcn_mask_head._do_paste_mask' function.

Args:
det_bboxes (np.ndarray): Bbox of shape [num_det, 4]
det_masks (np.ndarray): Masks of shape [num_det, 28, 28].
det_bboxes (np.ndarray | Tensor): Bbox of shape [num_det, 4]
det_masks (np.ndarray | Tensor): Masks of shape [num_det, 28, 28].
img_w (int): Width of the original image.
img_h (int): Height of the original image.
mask_thr_binary (float): The threshold for the mask.
device :(str): The device type.

Returns:
np.ndarray: masks of shape [N, num_det, img_h, img_w].
torch.Tensor: masks of shape [N, num_det, img_h, img_w].
"""
masks = det_masks
bboxes = det_bboxes

device = torch.device(device)
num_det = bboxes.shape[0]
# Skip postprocessing if no detections are found.
if num_det == 0:
return np.zeros((0, img_h, img_w))
return torch.zeros(
0, img_h, img_w, dtype=torch.float32, device=device)

if isinstance(masks, np.ndarray):
masks = torch.tensor(masks, device=torch.device(device))
bboxes = torch.tensor(bboxes, device=torch.device(device))
masks = torch.tensor(masks, device=device)
bboxes = torch.tensor(bboxes, device=device)

masks = masks.to(device)
bboxes = bboxes.to(device)

result_masks = []
for bbox, mask in zip(bboxes, masks):
Expand All @@ -150,15 +153,9 @@ def postprocessing_masks(det_bboxes: np.ndarray,
x1_int, y1_int = img_w, img_h

img_y = torch.arange(
y0_int,
y1_int,
dtype=torch.float32,
device=torch.device(device)) + 0.5
y0_int, y1_int, dtype=torch.float32, device=device) + 0.5
img_x = torch.arange(
x0_int,
x1_int,
dtype=torch.float32,
device=torch.device(device)) + 0.5
x0_int, x1_int, dtype=torch.float32, device=device) + 0.5
x0, y0, x1, y1 = bbox

img_y = (img_y - y0) / (y1 - y0) * 2 - 1
Expand Down Expand Up @@ -208,15 +205,13 @@ def forward(self, img: Sequence[torch.Tensor], img_metas: Sequence[dict],
rescale = kwargs.get('rescale', True)
for i in range(batch_size):
dets, labels = batch_dets[i], batch_labels[i]
dets = dets.to(device=torch.device(self.device))
if rescale:
scale_factor = img_metas[i]['scale_factor']

if isinstance(scale_factor, (list, tuple, np.ndarray)):
assert len(scale_factor) == 4
scale_factor = np.array(scale_factor)[None, :] # [1,4]
scale_factor = torch.from_numpy(scale_factor).to(
device=torch.device(self.device))
scale_factor = torch.from_numpy(scale_factor).to(dets)
dets[:, :4] /= scale_factor

if 'border' in img_metas[i]:
Expand Down Expand Up @@ -255,6 +250,8 @@ def forward(self, img: Sequence[torch.Tensor], img_metas: Sequence[dict],
masks = masks.squeeze(0)
if masks.dtype != bool:
masks = masks >= 0.5
# aligned with mmdet to easily convert to numpy
masks = masks.cpu()
segms_results = [[] for _ in range(len(self.CLASSES))]
for j in range(len(dets)):
segms_results[labels[j]].append(masks[j])
Expand Down Expand Up @@ -600,23 +597,21 @@ def forward_test(self, imgs: torch.Tensor, *args, **kwargs) -> List:
imgs (torch.Tensor): Input image(s) in [N x C x H x W] format.

Returns:
list[np.ndarray]: dets of shape [N, num_det, 5] and
list[torch.Tensor]: dets of shape [N, num_det, 5] and
class labels of shape [N, num_det].
"""
_, _, H, W = imgs.shape
outputs = self.wrapper({self.input_name: imgs})
for key, item in outputs.items():
if item is None:
return [np.zeros((1, 0, 5)), np.zeros((1, 0))]
return torch.zeros(1, 0, 5), torch.zeros(1, 0)
out = self.wrapper.output_to_list(outputs)[0]
labels = out[:, :, 0] - 1
scales = torch.tensor([W, H, W, H]).reshape(1, 1, 4)
scales = torch.tensor([W, H, W, H]).reshape(1, 1, 4).to(out)
scores = out[:, :, 1:2]
boxes = out[:, :, 2:6] * scales
dets = torch.cat([boxes, scores], dim=2)
dets = dets.detach().cpu().numpy()
labels = labels.detach().cpu().numpy()
return [dets, labels]
return dets, labels


@__BACKEND_MODEL.register_module('sdk')
Expand Down