Installation

Contents

Installation#

``uv`` is the recommended way to install MASE. It handles Python version pinning, dependency locking, and editable installs automatically. Follow the uv guide to get started in a few commands.

Docker is available as an alternative for fully isolated environments. A separate guide is provided for students setting up MASE for coursework.

Import a model#

To import a model into MASE and use all its features, the following options are available.

  • Generate a MaseGraph from a torch.nn.Module instance.

    • Since the MASE IR is based on Torch FX, symbolic tracing limitations apply to this method, namely models with control flow cannot be traced (see documentation).

    • Exising Pytorch models can be patched to remove control and run symbolic tracing (see here for examples).

import torch.nn as nn
from chop import MaseGraph

class MyModel(nn.Module):
    def __init__(self):
        ...

    def forward(self):
        ...

model = MyModel()
mg = MaseGraph(model)
Contents