Multilingual sentiment in NLP - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to load the multilingual sentiment dataset using Hugging Face datasets.

NLP

from datasets import load_dataset

dataset = load_dataset([1], 'multilingual')

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

A'imdb'

B'squad'

C'glue'

D'amazon_reviews_multi'

Attempts:

3 left

💡 Hint

Common Mistakes

Choosing 'imdb' which is English only.

Using 'glue' which is for English language understanding tasks.

Selecting 'squad' which is for question answering.

The 'amazon_reviews_multi' dataset contains multilingual sentiment reviews suitable for sentiment analysis tasks.

2fill in blank

medium

Complete the code to tokenize the input text for multilingual sentiment analysis using a pretrained tokenizer.

NLP

from transformers import AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained([1])
tokens = tokenizer(text, padding=True, truncation=True, return_tensors='pt')

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

A'bert-base-uncased'

B'xlm-roberta-base'

C'distilbert-base-uncased'

D'roberta-base'

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'bert-base-uncased' which is English only.

Choosing 'distilbert-base-uncased' which is English only.

Selecting 'roberta-base' which is English only.

The 'xlm-roberta-base' tokenizer supports multiple languages, making it suitable for multilingual sentiment tasks.

3fill in blank

hard

Fix the error in the model loading code for multilingual sentiment classification.

NLP

from transformers import AutoModelForSequenceClassification

model = AutoModelForSequenceClassification.from_pretrained([1], num_labels=3)

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

A'bert-base-multilingual-cased'

B'xlm-roberta-base'

C'bert-base-uncased'

D'distilbert-base-uncased'

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'bert-base-uncased' which is English only.

Choosing 'xlm-roberta-base' without fine-tuning for classification.

Selecting 'distilbert-base-uncased' which is English only.

The 'bert-base-multilingual-cased' model supports multiple languages and can be used for sequence classification with specified labels.

4fill in blank

hard

Fill both blanks to create a dictionary comprehension that maps each review text to its sentiment label.

NLP

sentiment_dict = { [1]: [2] for example in dataset['train'] }

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

Aexample['review_body']

Bexample['star_rating']

Cexample['text']

Dexample['sentiment']

Attempts:

3 left

💡 Hint

Common Mistakes

Choosing 'text' which does not exist in the dataset.

Using 'sentiment' which does not exist in this dataset.

The dataset uses 'review_body' for review content and 'star_rating' for sentiment (1-5 stars), so the dictionary maps text to star rating.

5fill in blank

hard

Fill all three blanks to compute accuracy of the model predictions.

NLP

correct = sum(1 for pred, true in zip(predictions, labels) if pred [1] true)
accuracy = correct [2] len(labels)
print(f'Accuracy: {accuracy:.2f}')

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

A==

B/

C*

D!=

Attempts:

3 left

💡 Hint

Common Mistakes

Using '!=' instead of '==' causing wrong accuracy calculation.

Multiplying instead of dividing for accuracy.

To compute accuracy, count where prediction equals true label, then divide by total labels.