TorchServe setup in PyTorch - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to import the TorchServe handler module.

PyTorch

from ts.torch_handler import [1]

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

Abase_handler

Btorchserve

Cmodel_handler

Dserve

Attempts:

3 left

💡 Hint

Common Mistakes

Trying to import torchserve or serve which are not modules in ts.torch_handler.

Using model_handler which is not a standard import.

The base_handler module is imported from ts.torch_handler to create custom handlers in TorchServe.

2fill in blank

medium

Complete the command to start TorchServe with a model store directory.

PyTorch

torchserve --start --model-store [1] --models my_model.mar

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

A/tmp/model_store

B/models

C/home/user/models

D/var/model

Attempts:

3 left

💡 Hint

Common Mistakes

Using a directory that does not exist or is not accessible.

Omitting the --model-store option.

The model store directory is specified with --model-store. /tmp/model_store is a common default directory.

3fill in blank

hard

Fix the error in the handler class definition by completing the base class name.

PyTorch

class MyHandler([1]):
    def __init__(self):
        super().__init__()

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

AHandlerBase

BBaseHandler

CTorchHandler

DModelHandler

Attempts:

3 left

💡 Hint

Common Mistakes

Using incorrect base class names like HandlerBase or TorchHandler.

Forgetting to call super().__init__() in the constructor.

Custom handlers inherit from BaseHandler class provided by TorchServe.

4fill in blank

hard

Fill both blanks to create a dictionary mapping model names to their .mar files for TorchServe registration.

PyTorch

models = [1]: '[2].mar' for [1] in ['resnet18', 'bert']}

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

Amodel_name

Bmodel

Cresnet18

Dbert

Attempts:

3 left

💡 Hint

Common Mistakes

Using fixed strings instead of variables in the comprehension.

Mixing variable names between key and value.

The dictionary comprehension uses model_name as the variable and key, and the value is model_name with .mar extension.

5fill in blank

hard

Fill all three blanks to define a TorchServe model archive creation command with model file, handler, and export path.

PyTorch

torch-model-archiver --model-name [1] --version 1.0 --serialized-file [2] --handler [3] --export-path model_store

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

Aresnet18

Bresnet18.pth

Cimage_classifier

Dbert

Ebert.pth

Ftext_classifier

Attempts:

3 left

💡 Hint

Common Mistakes

Mixing model names and handlers from different model types.

Using incorrect file names or handler names.

The model archive command uses the model name 'resnet18', the serialized file 'resnet18.pth', and the handler 'image_classifier'.