# Copyright (C) 2023-2024 Intel Corporation
# SPDX-License-Identifier: Apache-2.0
"""Class definition for classification model entity used in OTX."""
from __future__ import annotations
from typing import TYPE_CHECKING, Any, Literal
import numpy as np
import torch
from torch import Tensor
from otx.core.data.entity.base import OTXBatchLossEntity
from otx.core.data.entity.classification import (
HlabelClsBatchDataEntity,
HlabelClsBatchPredEntity,
MulticlassClsBatchDataEntity,
MulticlassClsBatchPredEntity,
MultilabelClsBatchDataEntity,
MultilabelClsBatchPredEntity,
)
from otx.core.exporter.base import OTXModelExporter
from otx.core.exporter.native import OTXNativeModelExporter
from otx.core.metrics import MetricInput
from otx.core.metrics.accuracy import (
HLabelClsMetricCallable,
MultiClassClsMetricCallable,
MultiLabelClsMetricCallable,
)
from otx.core.model.base import DefaultOptimizerCallable, DefaultSchedulerCallable, OTXModel, OVModel
from otx.core.schedulers import LRSchedulerListCallable
from otx.core.types.export import TaskLevelExportParameters
from otx.core.types.label import HLabelInfo, LabelInfo, LabelInfoTypes
from otx.core.types.task import OTXTrainType
if TYPE_CHECKING:
from lightning.pytorch.cli import LRSchedulerCallable, OptimizerCallable
from model_api.models.utils import ClassificationResult
from otx.core.metrics import MetricCallable
[docs]
class OTXMulticlassClsModel(OTXModel[MulticlassClsBatchDataEntity, MulticlassClsBatchPredEntity]):
"""Base class for the classification models used in OTX."""
def __init__(
self,
label_info: LabelInfoTypes,
input_size: tuple[int, int],
optimizer: OptimizerCallable = DefaultOptimizerCallable,
scheduler: LRSchedulerCallable | LRSchedulerListCallable = DefaultSchedulerCallable,
metric: MetricCallable = MultiClassClsMetricCallable,
torch_compile: bool = False,
train_type: Literal[OTXTrainType.SUPERVISED, OTXTrainType.SEMI_SUPERVISED] = OTXTrainType.SUPERVISED,
) -> None:
super().__init__(
label_info=label_info,
input_size=input_size,
optimizer=optimizer,
scheduler=scheduler,
metric=metric,
torch_compile=torch_compile,
train_type=train_type,
)
self.input_size: tuple[int, int]
def _customize_inputs(self, inputs: MulticlassClsBatchDataEntity) -> dict[str, Any]:
if self.training:
mode = "loss"
elif self.explain_mode:
mode = "explain"
else:
mode = "predict"
if self.train_type == OTXTrainType.SEMI_SUPERVISED and isinstance(inputs, dict):
# When used with an unlabeled dataset, it comes in as a dict.
images = {key: inputs[key].images for key in inputs}
labels = {key: torch.cat(inputs[key].labels, dim=0) for key in inputs}
imgs_info = {key: inputs[key].imgs_info for key in inputs}
return {
"images": images,
"labels": labels,
"imgs_info": imgs_info,
"mode": mode,
}
return {
"images": inputs.stacked_images,
"labels": torch.cat(inputs.labels, dim=0),
"imgs_info": inputs.imgs_info,
"mode": mode,
}
def _customize_outputs(
self,
outputs: Any, # noqa: ANN401
inputs: MulticlassClsBatchDataEntity,
) -> MulticlassClsBatchPredEntity | OTXBatchLossEntity:
if self.training:
return OTXBatchLossEntity(loss=outputs)
if self.explain_mode:
return MulticlassClsBatchPredEntity(
batch_size=inputs.batch_size,
images=inputs.images,
imgs_info=inputs.imgs_info,
scores=outputs["scores"],
labels=outputs["labels"],
saliency_map=outputs["saliency_map"],
feature_vector=outputs["feature_vector"],
)
# To list, batch-wise
logits = outputs if isinstance(outputs, torch.Tensor) else outputs["logits"]
scores = torch.unbind(logits, 0)
preds = logits.argmax(-1, keepdim=True).unbind(0)
return MulticlassClsBatchPredEntity(
batch_size=inputs.batch_size,
images=inputs.stacked_images,
imgs_info=inputs.imgs_info,
scores=scores,
labels=preds,
)
[docs]
def training_step(self, batch: MulticlassClsBatchDataEntity, batch_idx: int) -> Tensor:
"""Performs a single training step on a batch of data."""
loss = super().training_step(batch, batch_idx)
# Collect metrics related to Semi-SL Training.
if self.train_type == OTXTrainType.SEMI_SUPERVISED:
if hasattr(self.model, "unlabeled_coef"):
self.log(
"train/unlabeled_coef",
self.model.unlabeled_coef,
on_step=True,
on_epoch=False,
prog_bar=True,
)
if hasattr(self.model.head, "num_pseudo_label"):
self.log(
"train/num_pseudo_label",
self.model.head.num_pseudo_label,
on_step=True,
on_epoch=False,
prog_bar=True,
)
return loss
@property
def _export_parameters(self) -> TaskLevelExportParameters:
"""Defines parameters required to export a particular model implementation."""
return super()._export_parameters.wrap(
model_type="Classification",
task_type="classification",
multilabel=False,
hierarchical=False,
output_raw_scores=True,
)
@property
def _exporter(self) -> OTXModelExporter:
"""Creates OTXModelExporter object that can export the model."""
return OTXNativeModelExporter(
task_level_export_parameters=self._export_parameters,
input_size=(1, 3, *self.input_size),
mean=(123.675, 116.28, 103.53),
std=(58.395, 57.12, 57.375),
resize_mode="standard",
pad_value=0,
swap_rgb=False,
via_onnx=False,
onnx_export_configuration=None,
output_names=["logits", "feature_vector", "saliency_map"] if self.explain_mode else None,
)
def _convert_pred_entity_to_compute_metric(
self,
preds: MulticlassClsBatchPredEntity,
inputs: MulticlassClsBatchDataEntity,
) -> MetricInput:
pred = torch.tensor(preds.labels)
target = torch.tensor(inputs.labels)
return {
"preds": pred,
"target": target,
}
def _reset_prediction_layer(self, num_classes: int) -> None:
return
[docs]
def forward_for_tracing(self, image: Tensor) -> Tensor | dict[str, Tensor]:
"""Model forward function used for the model tracing during model exportation."""
return self.model(images=image)
### NOTE, currently, although we've made the separate Multi-cls, Multi-label classes
### It'll be integrated after H-label classification integration with more advanced design.
[docs]
class OTXMultilabelClsModel(OTXModel[MultilabelClsBatchDataEntity, MultilabelClsBatchPredEntity]):
"""Multi-label classification models used in OTX."""
def __init__(
self,
label_info: LabelInfoTypes,
input_size: tuple[int, int],
optimizer: OptimizerCallable = DefaultOptimizerCallable,
scheduler: LRSchedulerCallable | LRSchedulerListCallable = DefaultSchedulerCallable,
metric: MetricCallable = MultiLabelClsMetricCallable,
torch_compile: bool = False,
) -> None:
super().__init__(
label_info=label_info,
input_size=input_size,
optimizer=optimizer,
scheduler=scheduler,
metric=metric,
torch_compile=torch_compile,
)
self.input_size: tuple[int, int]
def _customize_inputs(self, inputs: MultilabelClsBatchDataEntity) -> dict[str, Any]:
if self.training:
mode = "loss"
elif self.explain_mode:
mode = "explain"
else:
mode = "predict"
return {
"images": inputs.stacked_images,
"labels": torch.stack(inputs.labels),
"imgs_info": inputs.imgs_info,
"mode": mode,
}
def _customize_outputs(
self,
outputs: Any, # noqa: ANN401
inputs: MultilabelClsBatchDataEntity,
) -> MultilabelClsBatchPredEntity | OTXBatchLossEntity:
if self.training:
return OTXBatchLossEntity(loss=outputs)
if self.explain_mode:
return MultilabelClsBatchPredEntity(
batch_size=inputs.batch_size,
images=inputs.images,
imgs_info=inputs.imgs_info,
scores=outputs["scores"],
labels=outputs["labels"],
saliency_map=outputs["saliency_map"],
feature_vector=outputs["feature_vector"],
)
# To list, batch-wise
logits = outputs if isinstance(outputs, torch.Tensor) else outputs["logits"]
scores = torch.unbind(logits, 0)
return MultilabelClsBatchPredEntity(
batch_size=inputs.batch_size,
images=inputs.images,
imgs_info=inputs.imgs_info,
scores=scores,
labels=logits.argmax(-1, keepdim=True).unbind(0),
)
@property
def _export_parameters(self) -> TaskLevelExportParameters:
"""Defines parameters required to export a particular model implementation."""
return super()._export_parameters.wrap(
model_type="Classification",
task_type="classification",
multilabel=True,
hierarchical=False,
confidence_threshold=0.5,
output_raw_scores=True,
)
@property
def _exporter(self) -> OTXModelExporter:
"""Creates OTXModelExporter object that can export the model."""
return OTXNativeModelExporter(
task_level_export_parameters=self._export_parameters,
input_size=(1, 3, *self.input_size),
mean=(123.675, 116.28, 103.53),
std=(58.395, 57.12, 57.375),
resize_mode="standard",
pad_value=0,
swap_rgb=False,
via_onnx=False,
onnx_export_configuration=None,
output_names=["logits", "feature_vector", "saliency_map"] if self.explain_mode else None,
)
def _convert_pred_entity_to_compute_metric(
self,
preds: MultilabelClsBatchPredEntity,
inputs: MultilabelClsBatchDataEntity,
) -> MetricInput:
return {
"preds": torch.stack(preds.scores),
"target": torch.stack(inputs.labels),
}
[docs]
def forward_for_tracing(self, image: Tensor) -> Tensor | dict[str, Tensor]:
"""Model forward function used for the model tracing during model exportation."""
return self.model.forward(image)
[docs]
class OTXHlabelClsModel(OTXModel[HlabelClsBatchDataEntity, HlabelClsBatchPredEntity]):
"""H-label classification models used in OTX."""
label_info: HLabelInfo
def __init__(
self,
label_info: HLabelInfo,
input_size: tuple[int, int],
optimizer: OptimizerCallable = DefaultOptimizerCallable,
scheduler: LRSchedulerCallable | LRSchedulerListCallable = DefaultSchedulerCallable,
metric: MetricCallable = HLabelClsMetricCallable,
torch_compile: bool = False,
) -> None:
super().__init__(
label_info=label_info,
input_size=input_size,
optimizer=optimizer,
scheduler=scheduler,
metric=metric,
torch_compile=torch_compile,
)
self.input_size: tuple[int, int]
def _customize_inputs(self, inputs: HlabelClsBatchDataEntity) -> dict[str, Any]:
if self.training:
mode = "loss"
elif self.explain_mode:
mode = "explain"
else:
mode = "predict"
return {
"images": inputs.stacked_images,
"labels": torch.stack(inputs.labels),
"imgs_info": inputs.imgs_info,
"mode": mode,
}
def _customize_outputs(
self,
outputs: Any, # noqa: ANN401
inputs: HlabelClsBatchDataEntity,
) -> HlabelClsBatchPredEntity | OTXBatchLossEntity:
if self.training:
return OTXBatchLossEntity(loss=outputs)
# To list, batch-wise
if isinstance(outputs, dict):
scores = outputs["scores"]
labels = outputs["labels"]
else:
scores = outputs
labels = outputs.argmax(-1, keepdim=True)
if self.explain_mode:
return HlabelClsBatchPredEntity(
batch_size=inputs.batch_size,
images=inputs.images,
imgs_info=inputs.imgs_info,
scores=scores,
labels=labels,
saliency_map=outputs["saliency_map"],
feature_vector=outputs["feature_vector"],
)
return HlabelClsBatchPredEntity(
batch_size=inputs.batch_size,
images=inputs.images,
imgs_info=inputs.imgs_info,
scores=scores,
labels=labels,
)
@property
def _export_parameters(self) -> TaskLevelExportParameters:
"""Defines parameters required to export a particular model implementation."""
return super()._export_parameters.wrap(
model_type="Classification",
task_type="classification",
multilabel=False,
hierarchical=True,
confidence_threshold=0.5,
output_raw_scores=True,
)
@property
def _exporter(self) -> OTXModelExporter:
"""Creates OTXModelExporter object that can export the model."""
return OTXNativeModelExporter(
task_level_export_parameters=self._export_parameters,
input_size=(1, 3, *self.input_size),
mean=(123.675, 116.28, 103.53),
std=(58.395, 57.12, 57.375),
resize_mode="standard",
pad_value=0,
swap_rgb=False,
via_onnx=False,
onnx_export_configuration=None,
output_names=["logits", "feature_vector", "saliency_map"] if self.explain_mode else None,
)
def _convert_pred_entity_to_compute_metric(
self,
preds: HlabelClsBatchPredEntity,
inputs: HlabelClsBatchDataEntity,
) -> MetricInput:
hlabel_info: HLabelInfo = self.label_info # type: ignore[assignment]
_labels = torch.stack(preds.labels) if isinstance(preds.labels, list) else preds.labels
_scores = torch.stack(preds.scores) if isinstance(preds.scores, list) else preds.scores
if hlabel_info.num_multilabel_classes > 0:
preds_multiclass = _labels[:, : hlabel_info.num_multiclass_heads]
preds_multilabel = _scores[:, hlabel_info.num_multiclass_heads :]
pred_result = torch.cat([preds_multiclass, preds_multilabel], dim=1)
else:
pred_result = _labels
return {
"preds": pred_result,
"target": torch.stack(inputs.labels),
}
@staticmethod
def _dispatch_label_info(label_info: LabelInfoTypes) -> LabelInfo:
if not isinstance(label_info, HLabelInfo):
raise TypeError(label_info)
return label_info
[docs]
def forward_for_tracing(self, image: Tensor) -> Tensor | dict[str, Tensor]:
"""Model forward function used for the model tracing during model exportation."""
return self.model(images=image)
[docs]
class OVMulticlassClassificationModel(
OVModel[MulticlassClsBatchDataEntity, MulticlassClsBatchPredEntity],
):
"""Classification model compatible for OpenVINO IR inference.
It can consume OpenVINO IR model path or model name from Intel OMZ repository
and create the OTX classification model compatible for OTX testing pipeline.
"""
def __init__(
self,
model_name: str,
model_type: str = "Classification",
async_inference: bool = True,
max_num_requests: int | None = None,
use_throughput_mode: bool = False,
model_api_configuration: dict[str, Any] | None = None,
metric: MetricCallable = MultiClassClsMetricCallable,
**kwargs,
) -> None:
super().__init__(
model_name=model_name,
model_type=model_type,
async_inference=async_inference,
max_num_requests=max_num_requests,
use_throughput_mode=use_throughput_mode,
model_api_configuration=model_api_configuration,
metric=metric,
)
def _customize_outputs(
self,
outputs: list[ClassificationResult],
inputs: MulticlassClsBatchDataEntity,
) -> MulticlassClsBatchPredEntity:
pred_labels = [torch.tensor(out.top_labels[0][0], dtype=torch.long, device=self.device) for out in outputs]
pred_scores = [torch.tensor(out.top_labels[0][2], device=self.device) for out in outputs]
if outputs and outputs[0].saliency_map.size != 0:
# Squeeze dim 4D => 3D, (1, num_classes, H, W) => (num_classes, H, W)
predicted_s_maps = [out.saliency_map[0] for out in outputs]
# Squeeze dim 2D => 1D, (1, internal_dim) => (internal_dim)
predicted_f_vectors = [out.feature_vector[0] for out in outputs]
return MulticlassClsBatchPredEntity(
batch_size=len(outputs),
images=inputs.images,
imgs_info=inputs.imgs_info,
scores=pred_scores,
labels=pred_labels,
saliency_map=predicted_s_maps,
feature_vector=predicted_f_vectors,
)
return MulticlassClsBatchPredEntity(
batch_size=len(outputs),
images=inputs.images,
imgs_info=inputs.imgs_info,
scores=pred_scores,
labels=pred_labels,
)
def _convert_pred_entity_to_compute_metric(
self,
preds: MulticlassClsBatchPredEntity,
inputs: MulticlassClsBatchDataEntity,
) -> MetricInput:
pred = torch.tensor(preds.labels)
target = torch.tensor(inputs.labels)
return {
"preds": pred,
"target": target,
}
[docs]
class OVMultilabelClassificationModel(OVModel[MultilabelClsBatchDataEntity, MultilabelClsBatchPredEntity]):
"""Multilabel classification model compatible for OpenVINO IR inference.
It can consume OpenVINO IR model path or model name from Intel OMZ repository
and create the OTX classification model compatible for OTX testing pipeline.
"""
def __init__(
self,
model_name: str,
model_type: str = "Classification",
async_inference: bool = True,
max_num_requests: int | None = None,
use_throughput_mode: bool = True,
model_api_configuration: dict[str, Any] | None = None,
metric: MetricCallable = MultiLabelClsMetricCallable,
**kwargs,
) -> None:
model_api_configuration = model_api_configuration if model_api_configuration else {}
model_api_configuration.update({"multilabel": True, "confidence_threshold": 0.0})
super().__init__(
model_name=model_name,
model_type=model_type,
async_inference=async_inference,
max_num_requests=max_num_requests,
use_throughput_mode=use_throughput_mode,
model_api_configuration=model_api_configuration,
metric=metric,
)
def _customize_outputs(
self,
outputs: list[ClassificationResult],
inputs: MultilabelClsBatchDataEntity,
) -> MultilabelClsBatchPredEntity:
pred_scores = [
torch.tensor([top_label[2] for top_label in out.top_labels], device=self.device) for out in outputs
]
if outputs and outputs[0].saliency_map.size != 0:
# Squeeze dim 4D => 3D, (1, num_classes, H, W) => (num_classes, H, W)
predicted_s_maps = [out.saliency_map[0] for out in outputs]
# Squeeze dim 2D => 1D, (1, internal_dim) => (internal_dim)
predicted_f_vectors = [out.feature_vector[0] for out in outputs]
return MultilabelClsBatchPredEntity(
batch_size=len(outputs),
images=inputs.images,
imgs_info=inputs.imgs_info,
scores=pred_scores,
labels=[],
saliency_map=predicted_s_maps,
feature_vector=predicted_f_vectors,
)
return MultilabelClsBatchPredEntity(
batch_size=len(outputs),
images=inputs.images,
imgs_info=inputs.imgs_info,
scores=pred_scores,
labels=[],
)
def _convert_pred_entity_to_compute_metric(
self,
preds: MultilabelClsBatchPredEntity,
inputs: MultilabelClsBatchDataEntity,
) -> MetricInput:
return {
"preds": torch.stack(preds.scores),
"target": torch.stack(inputs.labels),
}
[docs]
class OVHlabelClassificationModel(OVModel[HlabelClsBatchDataEntity, HlabelClsBatchPredEntity]):
"""Hierarchical classification model compatible for OpenVINO IR inference.
It can consume OpenVINO IR model path or model name from Intel OMZ repository
and create the OTX classification model compatible for OTX testing pipeline.
"""
def __init__(
self,
model_name: str,
model_type: str = "Classification",
async_inference: bool = True,
max_num_requests: int | None = None,
use_throughput_mode: bool = True,
model_api_configuration: dict[str, Any] | None = None,
metric: MetricCallable = HLabelClsMetricCallable,
**kwargs,
) -> None:
model_api_configuration = model_api_configuration if model_api_configuration else {}
model_api_configuration.update({"hierarchical": True, "output_raw_scores": True})
super().__init__(
model_name=model_name,
model_type=model_type,
async_inference=async_inference,
max_num_requests=max_num_requests,
use_throughput_mode=use_throughput_mode,
model_api_configuration=model_api_configuration,
metric=metric,
)
def _customize_outputs(
self,
outputs: list[ClassificationResult],
inputs: HlabelClsBatchDataEntity,
) -> HlabelClsBatchPredEntity:
all_pred_labels = []
all_pred_scores = []
for output in outputs:
logits = output.raw_scores
predicted_labels = []
predicted_scores = []
cls_heads_info = self.model.hierarchical_info["cls_heads_info"]
for i in range(cls_heads_info["num_multiclass_heads"]):
logits_begin, logits_end = cls_heads_info["head_idx_to_logits_range"][str(i)]
head_logits = logits[logits_begin:logits_end]
j = np.argmax(head_logits)
predicted_labels.append(j)
predicted_scores.append(head_logits[j])
if cls_heads_info["num_multilabel_classes"]:
logits_begin = cls_heads_info["num_single_label_classes"]
head_logits = logits[logits_begin:]
for i in range(head_logits.shape[0]):
predicted_scores.append(head_logits[i])
if head_logits[i] > self.model.confidence_threshold:
predicted_labels.append(1)
else:
predicted_labels.append(0)
all_pred_labels.append(torch.tensor(predicted_labels, dtype=torch.long, device=self.device))
all_pred_scores.append(torch.tensor(predicted_scores, device=self.device))
if outputs and outputs[0].saliency_map.size != 0:
# Squeeze dim 4D => 3D, (1, num_classes, H, W) => (num_classes, H, W)
predicted_s_maps = [out.saliency_map[0] for out in outputs]
# Squeeze dim 2D => 1D, (1, internal_dim) => (internal_dim)
predicted_f_vectors = [out.feature_vector[0] for out in outputs]
return HlabelClsBatchPredEntity(
batch_size=len(outputs),
images=inputs.images,
imgs_info=inputs.imgs_info,
scores=all_pred_scores,
labels=all_pred_labels,
saliency_map=predicted_s_maps,
feature_vector=predicted_f_vectors,
)
return HlabelClsBatchPredEntity(
batch_size=len(outputs),
images=inputs.images,
imgs_info=inputs.imgs_info,
scores=all_pred_scores,
labels=all_pred_labels,
)
def _convert_pred_entity_to_compute_metric(
self,
preds: HlabelClsBatchPredEntity,
inputs: HlabelClsBatchDataEntity,
) -> MetricInput:
cls_heads_info = self.model.hierarchical_info["cls_heads_info"]
num_multilabel_classes = cls_heads_info["num_multilabel_classes"]
num_multiclass_heads = cls_heads_info["num_multiclass_heads"]
if num_multilabel_classes > 0:
preds_multiclass = torch.stack(preds.labels)[:, :num_multiclass_heads]
preds_multilabel = torch.stack(preds.scores)[:, num_multiclass_heads:]
pred_result = torch.cat([preds_multiclass, preds_multilabel], dim=1)
else:
pred_result = torch.stack(preds.labels)
return {
"preds": pred_result,
"target": torch.stack(inputs.labels),
}
def _create_label_info_from_ov_ir(self) -> HLabelInfo:
ov_model = self.model.get_model()
if ov_model.has_rt_info(["model_info", "label_info"]):
serialized = ov_model.get_rt_info(["model_info", "label_info"]).value
return HLabelInfo.from_json(serialized)
msg = "Cannot construct LabelInfo from OpenVINO IR. Please check this model is trained by OTX."
raise ValueError(msg)
