morpheus.config.ConfigOnnxToTRT

class ConfigOnnxToTRT(input_model=None, output_model=None, batches=<factory>, seq_length=None, max_workspace_size=16000)[source]

Bases: morpheus.config.ConfigBase

Configuration class for the OnnxToTRT migration process.

Parameters
input_modelstr

Input ONNX model file.

output_modelstr

Output TensorRT model file.

batchestyping.List[typing.Tuple[int, int]]

Batches of input data.

seq_lengthint

Sequence length.

max_workspace_sizeint

Maximum workspace size, default 16000 MB.

Attributes
input_modelstr

Input ONNX model file.

output_modelstr

Output TensorRT model file.

seq_lengthint

Sequence length.

© Copyright 2023, NVIDIA. Last updated on Apr 11, 2023.