Hugging Face Transformers library in NLP - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to load a pre-trained BERT tokenizer from Hugging Face.

NLP

from transformers import [1]
tokenizer = [1].from_pretrained('bert-base-uncased')

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

ABertTokenizer

BBertModel

CAutoModel

DAutoTokenizer

Attempts:

3 left

💡 Hint

Common Mistakes

Using BertModel instead of BertTokenizer will cause errors because BertModel is for the model, not the tokenizer.

Using AutoModel instead of a tokenizer class.

The BertTokenizer class loads the tokenizer for BERT models. It converts text into tokens that the model understands.

2fill in blank

medium

Complete the code to tokenize a sentence using the tokenizer.

NLP

sentence = "Hello, how are you?"
tokens = tokenizer.[1](sentence, return_tensors='pt')

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

Aencode

Bencode_plus

Ctokenize

Dbatch_encode_plus

Attempts:

3 left

💡 Hint

Common Mistakes

Using tokenize returns only tokens as strings, not tensors.

Using encode returns token ids but not attention masks.

The encode_plus method tokenizes the input and returns tensors ready for the model, including attention masks.

3fill in blank

hard

Fix the error in the code to load a pre-trained BERT model for sequence classification.

NLP

from transformers import BertForSequenceClassification
model = BertForSequenceClassification.from_pretrained([1])

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

Abert-base

Bbert-base-cased

Cbert-base-uncased-cased

Dbert-base-uncased

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'bert-base-uncased-cased' is invalid because it mixes casing.

Using 'bert-base' is incomplete and will cause an error.

The correct model name is bert-base-uncased. It is a standard pre-trained BERT model without case sensitivity.

4fill in blank

hard

Fill both blanks to prepare inputs and get model outputs.

NLP

inputs = tokenizer([1], return_tensors=[2])
outputs = model(**inputs)

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

A"This is a test sentence."

Bpt

Ctf

D"Another example text."

Attempts:

3 left

💡 Hint

Common Mistakes

Passing a list instead of a string for a single sentence.

Using 'tf' when the model expects PyTorch tensors.

We pass a string sentence to the tokenizer and specify 'pt' to get PyTorch tensors for the model.

5fill in blank

hard

Fill all three blanks to extract predicted class from model logits.

NLP

import torch
logits = outputs.logits
predicted_class = torch.[1](logits, dim=[2]).[3]()

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

Aargmax

B1

Citem

Dmax

Attempts:

3 left

💡 Hint

Common Mistakes

Using max instead of argmax returns the max value, not the index.

Using dim=0 instead of dim=1 for batch outputs.

We use argmax over dimension 1 to get the predicted class index, then item() to get the Python number.