reload

R2026b

Reload PyTorch model from stored model loading properties

Since R2026b

Syntax

reload(model)

Description

reload(model) reloads the PyTorch^® model object from the stored model loading properties in model. Use this function to restore the Python^® model after terminating the Python environment, or to switch to a different model file while retaining the existing data-transfer and function configuration settings.

Examples

Reload Model After Terminating Python Environment

Reload the PyTorch model after the Python environment has been terminated and the underlying Python model object has been deleted.

Load a model.

model = pyTorchModel("model1OnPath.pt");
model.PythonObject

ans =
  Python MyModelClass1 with properties:

    training: 1

    MyModelClass1(
      (fc1): Linear(in_features=2, out_features=3, bias=True)
      (relu): ReLU()
      (fc2): Linear(in_features=3, out_features=5, bias=True)
    )

Terminate the Python environment. The Python model is deleted.

terminate(pyenv)
model.PythonObject

ans =
  handle to deleted MyModelClass1

Reload the Python model.

reload(model);
model.PythonObject

ans =
  Python MyModelClass1 with properties:

    training: 1

    MyModelClass1(
      (fc1): Linear(in_features=2, out_features=3, bias=True)
      (relu): ReLU()
      (fc2): Linear(in_features=3, out_features=5, bias=True)
    )

Preserve Settings While Changing Model File

Change the PyTorch model to a different one without specifying the data-transfer and added function settings again.

Load a model, specifying data-transfer settings.

model = pyTorchModel("mimo32_Full.pt",...
    NumInputs=3, ...
    InputDimensionOrder={[3 4 1 2],[2 3 1],[]},...
    InputNumDimensions={4 [] 3},...
    InputDataType={"float32","float32","float32"})

model =
  PyTorchModel with properties:

         PythonObject: [1x1 py.mimo32.MIMO32]
            ModelType: "torch.nn.Module"
             FileName: "mimo32_Full.pt"
           ModulePath: ""
           ModuleName: ""
    PyTorchFunctions: [1x1 struct]
            NumInputs: 3
  InputDimensionOrder: {[3 4 1 2]  [2 3 1]  []}
   InputNumDimensions: {[4]  []  [3]}
        InputDataType: ["float32"    "float32"    "float32"]
       InputKeyNames: [1x0 string]
 OutputDimensionOrder: {[]}
       OutputDataType: ""

Change the model filename and reload. The data-transfer settings are preserved.

model.FileName = "mimo32_PT26.pt2";
reload(model)

model =
  PyTorchModel with properties:

         PythonObject: [1x1 py.torch.fx.graph_module.GraphModule]
            ModelType: "ExportedProgram"
             FileName: "mimo32_PT26.pt2"
           ModulePath: ""
           ModuleName: ""
    PyTorchFunctions: [1x1 struct]
            NumInputs: 3
  InputDimensionOrder: {[]  []  []}
   InputNumDimensions: {[4]  [3]  [3]}
        InputDataType: ["float32"    "float32"    "float32"]
       InputKeyNames: [1x0 string]
 OutputDimensionOrder: {[]}
       OutputDataType: ""

Input Arguments

model — Reference to a PyTorch model

PyTorchModel object

Reference to a PyTorch model, specified as a PyTorchModel object. The model loading properties (FileName, ModulePath, ModuleName, or ConstructorCommand) stored in the object are used to reload the Python model.

Version History

Introduced in R2026b