Model serialization formats (pickle, ONNX, TorchScript) in MLOps - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to save a model using pickle.

MLOps

import pickle
with open('model.pkl', '[1]') as f:
    pickle.dump(model, f)

Drag options to blanks, or click blank then click option'

Ar

Brb

Cwb

Dw

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'r' or 'rb' mode which is for reading files.

Use 'wb' to write a binary file when saving a model with pickle.

2fill in blank

medium

Complete the code to export a PyTorch model to ONNX format.

MLOps

import torch
model.eval()
dummy_input = torch.randn(1, 3, 224, 224)
torch.onnx.export(model, dummy_input, '[1]')

Drag options to blanks, or click blank then click option'

Amodel.onnx

Bmodel.script

Cmodel.pkl

Dmodel.pt

Attempts:

3 left

💡 Hint

Common Mistakes

Saving ONNX model with '.pt' or '.pkl' extensions.

ONNX models are saved with the '.onnx' extension.

3fill in blank

hard

Fix the error in saving a TorchScript model.

MLOps

import torch
scripted_model = torch.jit.script(model)
scripted_model.[1]('model.pt')

Drag options to blanks, or click blank then click option'

Aexport

Bsave

Cdump

Dwrite

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'export' or 'dump' which are not valid for TorchScript.

TorchScript models use the 'save' method to write to disk.

4fill in blank

hard

Fill both blanks to load a pickled model correctly.

MLOps

import pickle
with open('model.pkl', '[1]') as f:
    model = pickle.[2](f)

Drag options to blanks, or click blank then click option'

Arb

Bwb

Cload

Ddump

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'wb' mode or pickle.dump() when loading.

To load a pickled model, open the file in 'rb' mode and use pickle.load().

5fill in blank

hard

Fill all three blanks to export a TorchScript model and save it.

MLOps

import torch
scripted = torch.jit.[1](model)
scripted.[2]('model_scripted.[3]')

Drag options to blanks, or click blank then click option'

Atrace

Bsave

Cpt

Dscript

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'trace' instead of 'script' or wrong file extensions.

Use torch.jit.script() to create a TorchScript model, then save() it with '.pt' extension.