metrics.categorical_accuracy.CategoricalAccuracy Class Reference

Collaboration diagram for metrics.categorical_accuracy.CategoricalAccuracy:

Public Member Functions
	__init__ (self)
	reset_state (self)
	update_state (self, outputs, expecteds)
	result (self)

Data Fields
	correct
	total
	axis

Detailed Description

Metric for computing categorical accuracy.

Definition at line 5 of file categorical_accuracy.py.

Constructor & Destructor Documentation

__init__()

metrics.categorical_accuracy.CategoricalAccuracy.__init__

(

self

)

Definition at line 9 of file categorical_accuracy.py.

    def __init__(self):
        self.correct = 0
        self.total = 0
        self.axis = 0
 

Member Function Documentation

reset_state()

metrics.categorical_accuracy.CategoricalAccuracy.reset_state

(

self

)

Reset the metric's state.

Reimplemented from metrics.metric.Metric.

Definition at line 14 of file categorical_accuracy.py.

    def reset_state(self):
        """
        Reset the metric's state.
        """
        self.correct = 0
        self.total = 0
 

References metrics.categorical_accuracy.CategoricalAccuracy.correct, and metrics.categorical_accuracy.CategoricalAccuracy.total.

result()

metrics.categorical_accuracy.CategoricalAccuracy.result

(

self

)

Compute and return the final metric value.
Returns:
    float: Metric value.

Reimplemented from metrics.metric.Metric.

Definition at line 48 of file categorical_accuracy.py.

    def result(self):
        """
        Compute and return the final metric value.
        Returns:
            float: Metric value.
        """
        if self.total == 0:
            return 0.0
        return self.correct / self.total

References metrics.categorical_accuracy.CategoricalAccuracy.correct, and metrics.categorical_accuracy.CategoricalAccuracy.total.

update_state()

metrics.categorical_accuracy.CategoricalAccuracy.update_state	(		self,
			outputs,
			expecteds
	)

Update the metric's state based on the outputs and expecteds.
Args:
    outputs (list of np.ndarray): List of model outputs for each output layer.
    expecteds (list of np.ndarray): List of expected ground truth values for each output layer.

Reimplemented from metrics.metric.Metric.

Definition at line 21 of file categorical_accuracy.py.

    def update_state(self, outputs, expecteds):
        """
        Update the metric's state based on the outputs and expecteds.
        Args:
            outputs (list of np.ndarray): List of model outputs for each output layer.
            expecteds (list of np.ndarray): List of expected ground truth values for each output layer.
        """
        if len(outputs) != len(expecteds):
            raise ValueError(
                "The number of outputs and expecteds must match. "
                f"Got {len(outputs)} outputs and {len(expecteds)} expecteds.")
 
        for output, expected in zip(outputs, expecteds):
            if output.shape[self.axis] != expected.shape[self.axis]:
                raise ValueError(
                    f"Output and expected shapes must match along the specified axis {self.axis}. "
                    f"Got output shape {output.shape} and expected shape {expected.shape}."
                )
 
            batch_size = output.shape[self.axis]
            for b in range(batch_size):
                output_idx = np.argmax(output[b])
                expected_idx = np.argmax(expected[b])
                if output_idx == expected_idx:
                    self.correct += 1
            self.total += batch_size
 

References luci_interpreter::kernels::CumSum.axis(), coco::ConcatF.axis(), ShapeQuery.axis(), luci_interpreter::ConcatenationParams.axis, luci_interpreter::GatherParams.axis, luci_interpreter::OneHotParams.axis, luci_interpreter::PackParams.axis, luci_interpreter::UnpackParams.axis, luci_interpreter::kernels::ArgMax.axis(), luci_interpreter::kernels::ExpandDims.axis(), luci_interpreter::kernels::Split.axis(), luci_interpreter::kernels::SplitV.axis(), loco::Permutation< Domain::DepthwiseFilter >.axis(), loco::Permutation< Domain::DepthwiseFilter >.axis(), loco::Permutation< Domain::Feature >.axis(), loco::Permutation< Domain::Feature >.axis(), loco::Permutation< Domain::Filter >.axis(), loco::Permutation< Domain::Filter >.axis(), loco::Permutation< Domain::Matrix >.axis(), loco::Permutation< Domain::Matrix >.axis(), luci::CircleBCQGather.axis(), luci::CircleConcatenation.axis(), luci::CircleGather.axis(), luci::CircleOneHot.axis(), luci::CirclePack.axis(), luci::CircleUnpack.axis(), moco::TFPack.axis(), luci::CircleCumSum.axis(), luci::CircleExpandDims.axis(), luci::CircleReverseV2.axis(), moco::TFConcatV2.axis(), loco::TensorTranspose::Perm.axis(), loco::TensorTranspose::Perm.axis(), locoex::TFLConcatenation.axis(), loco::Softmax< Domain::Tensor >.axis(), loco::BiasAdd< Domain::Tensor >.axis(), moco::TensorPackEnumerator.axis(), loco::TensorConcat.axis(), coco::ConcatF.axis(), locoex::TFLConcatenation.axis(), loco::TensorConcat.axis(), loco::Softmax< Domain::Tensor >.axis(), loco::BiasAdd< Domain::Tensor >.axis(), luci::CircleBCQGather.axis(), luci::CircleConcatenation.axis(), luci::CircleCumSum.axis(), luci::CircleExpandDims.axis(), luci::CircleGather.axis(), luci::CircleOneHot.axis(), luci::CirclePack.axis(), luci::CircleReverseV2.axis(), luci::CircleUnpack.axis(), moco::TFConcatV2.axis(), moco::TFPack.axis(), luci_interpreter_pal::MeanParams.axis, luci_interpreter_pal::ConcatenationParams.axis, onert_micro::core::ConcatenationParams.axis, nnfw::cker::SoftmaxParams.axis, nnfw::cker::PackParams.axis, nnfw::cker::UnpackParams.axis, nnfw::cker::ConcatenationParams.axis, nnfw::cker::GatherParams.axis, nnfw::cker::SplitParams.axis, nnfw::cker::SplitVParams.axis, metrics.categorical_accuracy.CategoricalAccuracy.axis, onert::ir::operation::BCQGather::Param.axis, onert::ir::operation::Concat::Param.axis, onert::ir::operation::Gather::Param.axis, onert::ir::operation::LogSoftmax::Param.axis, onert::ir::operation::OneHot::Param.axis, onert::ir::operation::Pack::Param.axis, onert::ir::operation::Unpack::Param.axis, onert::ir::train::CategoricalCrossentropyParam.axis, metrics.categorical_accuracy.CategoricalAccuracy.correct, and metrics.categorical_accuracy.CategoricalAccuracy.total.

Field Documentation

axis

metrics.categorical_accuracy.CategoricalAccuracy.axis

Definition at line 12 of file categorical_accuracy.py.

Referenced by metrics.categorical_accuracy.CategoricalAccuracy.update_state().

correct

metrics.categorical_accuracy.CategoricalAccuracy.correct

Definition at line 10 of file categorical_accuracy.py.

Referenced by metrics.categorical_accuracy.CategoricalAccuracy.reset_state(), metrics.categorical_accuracy.CategoricalAccuracy.result(), and metrics.categorical_accuracy.CategoricalAccuracy.update_state().

total

metrics.categorical_accuracy.CategoricalAccuracy.total

Definition at line 11 of file categorical_accuracy.py.

Referenced by metrics.categorical_accuracy.CategoricalAccuracy.reset_state(), metrics.categorical_accuracy.CategoricalAccuracy.result(), and metrics.categorical_accuracy.CategoricalAccuracy.update_state().

---

The documentation for this class was generated from the following file:

• runtime/onert/api/python/package/experimental/train/metrics/categorical_accuracy.py