Class ov::op::v14::RMSNorm#

class RMSNorm : public ov::op::Op#

Operator performing Root Mean Square Normalization.

Public Functions

RMSNorm(const Output<Node> &data, const Output<Node> &axes, double epsilson, const ov::element::Type &compute_type = ov::element::undefined)#

Constructs an RMSNorm operation without scaling.

Parameters:

  • data – Input tensor with data

  • axes – Axes for reduce mean calculation

  • eps – Epsilon for not dividing by zero while normalizing the value

  • compute_type – Precision for the internal computation, if undefined it’s the same as the input type

RMSNorm(const Output<Node> &data, const Output<Node> &axes, const Output<Node> &scale, double epsilson, const ov::element::Type &compute_type = ov::element::undefined)#

Constructs an RMSNorm operation with scaling.

Parameters:

  • data – Input tensor with data

  • axes – Axes for reduce mean calculation

  • scale – Scale values for weight

  • eps – Epsilon for not dividing by zero while normalizing the value

  • compute_type – Precision for the internal computation, if undefined it’s the same as the input type

virtual void validate_and_infer_types() override#

Verifies that attributes and inputs are consistent and computes output shapes and element types. Must be implemented by concrete child classes so that it can be run any number of times.

Throws if the node is invalid.