onnx_diagnostic.reference.torch_ops.sequence_ops

class onnx_diagnostic.reference.torch_ops.sequence_ops.ConcatFromSequence_11(node: NodeProto, version: int | None = None)

ConcatFromSequence

run(input_sequence: OpRunSequence) → OpRunTensor

Kernel implementation.

class onnx_diagnostic.reference.torch_ops.sequence_ops.OpRunOpSequence(node: NodeProto, version: int | None = None)

Ancestor for kernel using sequences.

class onnx_diagnostic.reference.torch_ops.sequence_ops.SequenceEmpty_11(node: NodeProto, version: int | None = None)

SqeuenceEmpty

run() → OpRunSequence

Kernel implementation.

class onnx_diagnostic.reference.torch_ops.sequence_ops.SequenceInsert_11(node: NodeProto, version: int | None = None)

SqeuenceInsert

run(input_sequence: OpRunSequence, tensor: OpRunTensor, position: OpRunTensor | None = None) → OpRunSequence

Kernel implementation.