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

Add hlabel parent resolver and integrate HLabelData + HLabelInfo --> HLabelInfo #3046

Merged
merged 18 commits into from
Mar 11, 2024
Merged

Add hlabel parent resolver and integrate HLabelData + HLabelInfo --> HLabelInfo #3046

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
18 commits
Select commit Hold shift + click to select a range
5bb1309
Initial commit
sungmanc Mar 6, 2024
b565e57
Add function: add ancestor
sungmanc Mar 6, 2024
c35b187
Merge branch 'v2' of https://github.com/openvinotoolkit/training_exte…
sungmanc Mar 6, 2024
a8e3c45
Integrate HLabelData and HLabelInfo
sungmanc Mar 6, 2024
942dda6
Add unit-test for _add_ancestor func
sungmanc Mar 6, 2024
ae4b093
Fix invalid names
sungmanc Mar 6, 2024
f423ba3
Fix precommit
sungmanc Mar 6, 2024
0163c8e
Fix hlabel{
sungmanc Mar 7, 2024
6254cb3
Add type checking
sungmanc Mar 7, 2024
80eaaf2
Merge branch 'v2' of https://github.com/openvinotoolkit/training_exte…
sungmanc Mar 8, 2024
9d4e1fe
enlarge the threshold
sungmanc Mar 8, 2024
ced502d
Fix intg test
sungmanc Mar 8, 2024
83f4295
Merge branch 'v2' into add-hlabel-parent-resolver
sungmanc Mar 8, 2024
cda84f9
Remove mistaken line
sungmanc Mar 8, 2024
80a8be7
Remove mistaken line
sungmanc Mar 8, 2024
68b363e
Merge branch 'v2' of https://github.com/openvinotoolkit/training_exte…
sungmanc Mar 8, 2024
1a47b04
Merge branch 'add-hlabel-parent-resolver' of https://github.com/sungm…
sungmanc Mar 8, 2024
471b7a9
Merge upstream
sungmanc Mar 11, 2024
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
26 changes: 13 additions & 13 deletions src/otx/algo/classification/heads/custom_hlabel_linear_cls_head.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,7 +14,7 @@
from torch import nn

if TYPE_CHECKING:
from otx.core.data.entity.classification import HLabelData
from otx.core.data.dataset.classification import HLabelInfo


@MODELS.register_module()
Expand Down Expand Up @@ -74,19 +74,19 @@ def forward(self, feats: tuple[torch.Tensor]) -> torch.Tensor:
pre_logits = self.pre_logits(feats)
return self.fc(pre_logits)

def set_hlabel_data(self, hlabel_data: HLabelData) -> None:
def set_hlabel_info(self, hlabel_info: HLabelInfo) -> None:
"""Set hlabel information."""
self.hlabel_data = hlabel_data
self.hlabel_info = hlabel_info

def _get_gt_label(self, data_samples: list[DataSample]) -> torch.Tensor:
"""Get gt labels from data samples."""
return torch.stack([data_sample.gt_label for data_sample in data_samples])

def _get_head_idx_to_logits_range(self, hlabel_data: HLabelData, idx: int) -> tuple[int, int]:
def _get_head_idx_to_logits_range(self, hlabel_info: HLabelInfo, idx: int) -> tuple[int, int]:
"""Get head_idx_to_logits_range information from hlabel information."""
return (
hlabel_data.head_idx_to_logits_range[str(idx)][0],
hlabel_data.head_idx_to_logits_range[str(idx)][1],
hlabel_info.head_idx_to_logits_range[str(idx)][0],
hlabel_info.head_idx_to_logits_range[str(idx)][1],
)

def loss(self, feats: tuple[torch.Tensor], data_samples: list[DataSample], **kwargs) -> dict:
Expand All @@ -113,9 +113,9 @@ def loss(self, feats: tuple[torch.Tensor], data_samples: list[DataSample], **kwa
# Multiclass loss
num_effective_heads_in_batch = 0 # consider the label removal case
for i in range(self.num_multiclass_heads):
if i not in self.hlabel_data.empty_multiclass_head_indices:
if i not in self.hlabel_info.empty_multiclass_head_indices:
head_gt = gt_labels[:, i]
logit_range = self._get_head_idx_to_logits_range(self.hlabel_data, i)
logit_range = self._get_head_idx_to_logits_range(self.hlabel_info, i)
head_logits = cls_scores[:, logit_range[0] : logit_range[1]]
valid_mask = head_gt >= 0

Expand All @@ -130,15 +130,15 @@ def loss(self, feats: tuple[torch.Tensor], data_samples: list[DataSample], **kwa

# Multilabel loss
if self.num_multilabel_classes > 0:
head_gt = gt_labels[:, self.hlabel_data.num_multiclass_heads :]
head_logits = cls_scores[:, self.hlabel_data.num_single_label_classes :]
head_gt = gt_labels[:, self.hlabel_info.num_multiclass_heads :]
head_logits = cls_scores[:, self.hlabel_info.num_single_label_classes :]
valid_mask = head_gt > 0
head_gt = head_gt[valid_mask]
if len(head_gt) > 0:
img_metas = [data_sample.metainfo for data_sample in data_samples]
head_logits = head_logits[valid_mask]
valid_label_mask = self.get_valid_label_mask(img_metas).to(head_logits.device)
valid_label_mask = valid_label_mask[:, self.hlabel_data.num_single_label_classes :]
valid_label_mask = valid_label_mask[:, self.hlabel_info.num_single_label_classes :]
valid_label_mask = valid_label_mask[valid_mask]
losses["loss"] += self.multilabel_loss(head_logits, head_gt, valid_label_mask=valid_label_mask)

Expand Down Expand Up @@ -183,7 +183,7 @@ def _get_predictions(
multiclass_pred_scores = []
multiclass_pred_labels = []
for i in range(self.num_multiclass_heads):
logit_range = self._get_head_idx_to_logits_range(self.hlabel_data, i)
logit_range = self._get_head_idx_to_logits_range(self.hlabel_info, i)
multiclass_logit = cls_scores[:, logit_range[0] : logit_range[1]]
multiclass_pred = torch.softmax(multiclass_logit, dim=1)
multiclass_pred_score, multiclass_pred_label = torch.max(multiclass_pred, dim=1)
Expand All @@ -195,7 +195,7 @@ def _get_predictions(
multiclass_pred_labels = torch.cat(multiclass_pred_labels, dim=1)

if self.num_multilabel_classes > 0:
multilabel_logits = cls_scores[:, self.hlabel_data.num_single_label_classes :]
multilabel_logits = cls_scores[:, self.hlabel_info.num_single_label_classes :]

multilabel_pred_scores = torch.sigmoid(multilabel_logits)
multilabel_pred_labels = (multilabel_pred_scores >= self.thr).int()
Expand Down
26 changes: 13 additions & 13 deletions src/otx/algo/classification/heads/custom_hlabel_non_linear_cls_head.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,7 +14,7 @@
from torch import nn

if TYPE_CHECKING:
from otx.core.data.entity.classification import HLabelData
from otx.core.data.dataset.classification import HLabelInfo


@MODELS.register_module()
Expand Down Expand Up @@ -105,19 +105,19 @@ def forward(self, feats: tuple[torch.Tensor]) -> torch.Tensor:
pre_logits = self.pre_logits(feats)
return self.classifier(pre_logits)

def set_hlabel_data(self, hlabel_data: HLabelData) -> None:
def set_hlabel_info(self, hlabel_info: HLabelInfo) -> None:
"""Set hlabel information."""
self.hlabel_data = hlabel_data
self.hlabel_info = hlabel_info

def _get_gt_label(self, data_samples: list[DataSample]) -> torch.Tensor:
"""Get gt labels from data samples."""
return torch.stack([data_sample.gt_label for data_sample in data_samples])

def _get_head_idx_to_logits_range(self, hlabel_data: HLabelData, idx: int) -> tuple[int, int]:
def _get_head_idx_to_logits_range(self, hlabel_info: HLabelInfo, idx: int) -> tuple[int, int]:
"""Get head_idx_to_logits_range information from hlabel information."""
return (
hlabel_data.head_idx_to_logits_range[str(idx)][0],
hlabel_data.head_idx_to_logits_range[str(idx)][1],
hlabel_info.head_idx_to_logits_range[str(idx)][0],
hlabel_info.head_idx_to_logits_range[str(idx)][1],
)

def loss(self, feats: tuple[torch.Tensor], data_samples: list[DataSample], **kwargs) -> dict:
Expand All @@ -144,9 +144,9 @@ def loss(self, feats: tuple[torch.Tensor], data_samples: list[DataSample], **kwa
# Multiclass loss
num_effective_heads_in_batch = 0 # consider the label removal case
for i in range(self.num_multiclass_heads):
if i not in self.hlabel_data.empty_multiclass_head_indices:
if i not in self.hlabel_info.empty_multiclass_head_indices:
head_gt = gt_labels[:, i]
logit_range = self._get_head_idx_to_logits_range(self.hlabel_data, i)
logit_range = self._get_head_idx_to_logits_range(self.hlabel_info, i)
head_logits = cls_scores[:, logit_range[0] : logit_range[1]]
valid_mask = head_gt >= 0

Expand All @@ -161,15 +161,15 @@ def loss(self, feats: tuple[torch.Tensor], data_samples: list[DataSample], **kwa

# Multilabel loss
if self.num_multilabel_classes > 0:
head_gt = gt_labels[:, self.hlabel_data.num_multiclass_heads :]
head_logits = cls_scores[:, self.hlabel_data.num_single_label_classes :]
head_gt = gt_labels[:, self.hlabel_info.num_multiclass_heads :]
head_logits = cls_scores[:, self.hlabel_info.num_single_label_classes :]
valid_mask = head_gt > 0
head_gt = head_gt[valid_mask]
if len(head_gt) > 0:
img_metas = [data_sample.metainfo for data_sample in data_samples]
head_logits = head_logits[valid_mask]
valid_label_mask = self.get_valid_label_mask(img_metas).to(head_logits.device)
valid_label_mask = valid_label_mask[:, self.hlabel_data.num_single_label_classes :]
valid_label_mask = valid_label_mask[:, self.hlabel_info.num_single_label_classes :]
valid_label_mask = valid_label_mask[valid_mask]
losses["loss"] += self.multilabel_loss(head_logits, head_gt, valid_label_mask=valid_label_mask)

Expand Down Expand Up @@ -214,7 +214,7 @@ def _get_predictions(
multiclass_pred_scores = []
multiclass_pred_labels = []
for i in range(self.num_multiclass_heads):
logit_range = self._get_head_idx_to_logits_range(self.hlabel_data, i)
logit_range = self._get_head_idx_to_logits_range(self.hlabel_info, i)
multiclass_logit = cls_scores[:, logit_range[0] : logit_range[1]]
multiclass_pred = torch.softmax(multiclass_logit, dim=1)
multiclass_pred_score, multiclass_pred_label = torch.max(multiclass_pred, dim=1)
Expand All @@ -226,7 +226,7 @@ def _get_predictions(
multiclass_pred_labels = torch.cat(multiclass_pred_labels, dim=1)

if self.num_multilabel_classes > 0:
multilabel_logits = cls_scores[:, self.hlabel_data.num_single_label_classes :]
multilabel_logits = cls_scores[:, self.hlabel_info.num_single_label_classes :]

multilabel_pred_scores = torch.sigmoid(multilabel_logits)
multilabel_pred_labels = (multilabel_pred_scores >= self.thr).int()
Expand Down
Loading
Loading