Skip to main content

max visualize

Generates a file you can use to visualize a model with Netron.app.

Synopsis

max visualize [options] <model path>

Description

Generates a .maxviz file from a given TorchScript, TensorFlow, or ONNX model, which you can then load into Netron.app. The maxviz file is a deconstructed version of the model before it is optimized in the MAX Engine, meaning that it represents the high-level graph layout using the "MO" dialect. For more detail, see the guide to visualize a model.

Options

Output

--output <PATH>, -o <PATH>

Directory or path for the output maxviz file. By default, the maxviz file is saved to the current directory with the same name as the input file (but with the maxviz extension). You can override this with just a directory name or a full path with a different filename. If a matching filename exists already, this will override it.

Compilation options

--input-data-schema <PATH>

Path to input data schema file (only used for compile-time shape specifications).

--custom-ops-path <PATH>

Path to custom ops.

Common options

--help, -h

Displays help information.

--version

Displays version information.