Class ov::frontend::SequenceMark

class SequenceMark : public ov::op::util::FrameworkNode

SequenceMark serves to mark places that require a sequence type propagation. This class represents list or tuple constructs in frameworks like PyTorch, ONNX, etc. It holds a collection of tensors that form a sequence.

Public Functions

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.

ov::OutputVector get_sequence() const

Returns all sequence elements of this SequenceMark, flattening any SequenceInsert / SequenceErase chain (with statically known positions) and unwrapping Identity passthrough. A directly nested SequenceMark input is kept as a single opaque element, preserving the index->element mapping expected by tuple/list unpack and getitem. An Insert/Erase with a non-constant position, or a producer that carries a sequence across a Loop/If boundary, is also treated as a single opaque element.