Support Matrix

Please find below information about tested models, used environment and libraries.

Verified Models

We have verified that the NVIDIA Triton Model Navigator Optimize works correctly for the following models.

Source	Model
NVIDIA DeepLearningExamples	ResNet50 PyT
NVIDIA DeepLearningExamples	EfficientNet PyT
NVIDIA DeepLearningExamples	EfficientNet TF2
NVIDIA DeepLearningExamples	BERT TF2
HuggingFace	GPT2 Jax
HuggingFace	GPT2 PyT
HuggingFace	GPT2 TF2
HuggingFace	DistilBERT PyT
HuggingFace	DistilGPT2 TF2
HuggingFace	Stable Diffusion
HuggingFace	Automatic Speech Recognition

Libraries and Packages

A set of component versions are imposed by the used NGC container. During testing, we have used 24.02 a container version that contains:

• PyTorch 2.3.0a0+ebedce2

• TensorFlow 2.15.0

• TensorRT 8.6.3

• Torch-TensorRT 2.0.0.dev0

• ONNX Runtime 1.17.1

• Polygraphy: 0.49.4

• GraphSurgeon: 0.4.6

• tf2onnx v1.16.1

• Other components versions depend on the used framework containers versions. See its support matrix for a detailed summary.