Files

inference

Update lockfiles to point to latest nightly version: 25.3.0.dev202504…

Apr 11, 2025

f5adc05 · Apr 11, 2025

Name		Name	Last commit message	Last commit date
parent directory ..
bert-c-torchscript		bert-c-torchscript	Update lockfiles to point to latest nightly version: 25.3.0.dev202504…	Apr 11, 2025
bert-mojo-torchscript		bert-mojo-torchscript	Update lockfiles to point to latest nightly version: 25.3.0.dev202504…	Apr 11, 2025
bert-python-torchscript		bert-python-torchscript	Update lockfiles to point to latest nightly version: 25.3.0.dev202504…	Apr 11, 2025
common/bert-torchscript		common/bert-torchscript	[Pipelines] Compile multi-model pipelines' models in parallel	Feb 27, 2025
resnet50-python-torchscript		resnet50-python-torchscript	Update lockfiles to point to latest nightly version: 25.3.0.dev202504…	Apr 11, 2025
stable-diffusion-python-onnx		stable-diffusion-python-onnx	[examples] Remove temporary need for `accelerate`	Apr 9, 2025
yolo-python-onnx		yolo-python-onnx	Update lockfiles to point to latest nightly version: 25.3.0.dev202504…	Apr 11, 2025
README.md		README.md	[Pipelines] Compile multi-model pipelines' models in parallel	Feb 27, 2025

README.md

PyTorch and ONNX inference in MAX

MAX can accelerate the inference of existing PyTorch or ONNX models. These examples show several common PyTorch or ONNX models running in MAX through the Mojo, Python, and C APIs:

PyTorch (via TorchScript)

BERT, with the Mojo API, with the Python API, and with the C API
ResNet50 with the Python API

ONNX

Stable Diffusion, with the Mojo API and with the Python API
YOLOv8 with the Python API