QA with Hugging Face pipeline in NLP - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to create a question answering pipeline using Hugging Face.

NLP

from transformers import pipeline
qa_pipeline = pipeline([1])

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

A"question-answering"

B"translation"

C"sentiment-analysis"

D"text-generation"

Attempts:

3 left

💡 Hint

Common Mistakes

Using the wrong pipeline type like "text-generation" or "sentiment-analysis".

The pipeline type for question answering is "question-answering".

2fill in blank

medium

Complete the code to provide the context and question to the QA pipeline.

NLP

result = qa_pipeline({"question": [1], "context": "Hugging Face makes machine learning easy."})

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

A'Where is the moon?'

B'How to train a model?'

C'What is Python?'

D'What is Hugging Face?'

Attempts:

3 left

💡 Hint

Common Mistakes

Asking questions unrelated to the context, which leads to wrong answers.

The question should relate to the context. "What is Hugging Face?" fits the context about Hugging Face.

3fill in blank

hard

Fix the error in accessing the answer from the result dictionary.

NLP

print(result[[1]])

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

A"text"

B"answer"

C"output"

D"result"

Attempts:

3 left

💡 Hint

Common Mistakes

Using incorrect keys like "result" or "text" which do not exist in the output.

The QA pipeline returns a dictionary with the key "answer" containing the predicted answer.

4fill in blank

hard

Fill both blanks to create a QA pipeline with a specific model and tokenizer.

NLP

qa_pipeline = pipeline("question-answering", model=[1], tokenizer=[2])

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

A"distilbert-base-uncased-distilled-squad"

B"bert-base-uncased"

C"gpt2"

D"roberta-base"

Attempts:

3 left

💡 Hint

Common Mistakes

Using different names for model and tokenizer causing errors.

Choosing models not trained for QA like "gpt2".

The model and tokenizer names must match. "distilbert-base-uncased-distilled-squad" is a common QA model and tokenizer.

5fill in blank

hard

Fill all three blanks to extract the answer, score, and start position from the QA result.

NLP

answer = result[[1]]
score = result[[2]]
start = result[[3]]

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

A"answer"

B"score"

C"start"

D"end"

Attempts:

3 left

💡 Hint

Common Mistakes

Using wrong keys like "end" instead of "start" for the start position.

The result dictionary keys are "answer" for the text, "score" for confidence, and "start" for the answer start index.