Class InferenceMemoryFIL

class InferenceMemoryFIL : public morpheus::InferenceMemory

This is a container class for data that needs to be submitted to the inference server for FIL category usecases.

Public Functions

InferenceMemoryFIL(TensorIndex count, TensorObject &&input__0, TensorObject &&seq_ids)

Construct a new Inference Memory FIL object.

Parameters
  • count – : Message count in inference memory object

  • input__0 – : Inference input

  • seq_ids – : Ids used to index from an inference input to a message. Necessary since there can be more inference inputs than messages (i.e., if some messages get broken into multiple inference requests)

const TensorObject &get_input__0() const

Returns the ‘input__0’ tensor, throws a std::runtime_error if it does not exist.

Throws

std::runtime_error

Returns

const TensorObject&

const TensorObject &get_seq_ids() const

Returns the ‘seq_ids’ tensor, throws a std::runtime_error if it does not exist.

Throws

std::runtime_error

Returns

const TensorObject&

void set_input__0(TensorObject &&input__0)

Sets a tensor named ‘input__0’.

Parameters

input__0

Throws

std::length_error – If the number of rows in input__0 does not match count.

void set_seq_ids(TensorObject &&seq_ids)

Sets a tensor named ‘seq_ids’.

Parameters

seq_ids

Throws

std::length_error – If the number of rows in seq_ids does not match count.

Previous Class InferenceMemory
Next Class InferenceMemoryNLP
© Copyright 2023, NVIDIA. Last updated on Feb 2, 2024.