1. Question 1

In the lectures, what is the name of the layer used to generate the vocabulary?

• ❌ Tokenizer

• ❌ TextTokenizer

• ❌ WordTokenizer

• ✅ TextVectorization

Explanation: TensorFlow’s TextVectorization layer builds the vocabulary and vectorizes text.

---

2. Question 2

Once you have generated a vocabulary, how do you encode a sentence into an integer sequence?

• ✅ Pass the string to the adapted TextVectorization layer.

• ❌ get_vocabulary()

• ❌ texts_to_tokens()

• ❌ texts_to_sequences()

Explanation: The TextVectorization layer itself is callable.
Usage:

---

3. Question 3

How do you handle sequences of different lengths?

• ✅ Use the pad_sequences function from tf.keras.utils

• ❌ pad_sequences method inside TextVectorization

• ❌ pad_sequences property on input layer

• ❌ dynamic_length input layer

Explanation: Use tf.keras.utils.pad_sequences() to ensure equal sequence length.

---

4. Question 4

What happens when encoding a word not in the vocabulary?

• ❌ Sequencing ends

• ❌ Replaced by most common token

• ✅ An out-of-vocabulary token (OOV token) is used

• ❌ Replaced by zero

Explanation: OOV token (usually index 1) represents unknown words.

---

5. Question 5

How to pad at the end of a sequence?

• ✅ padding='post' in pad_sequences

• ❌ padding=’after’

• ❌ padding method with ‘after’

• ❌ padding method with ‘post’

Explanation:

---

6. Question 6

Convert a list of strings into integer sequences?

• ❌ vectorize_layer.fit(sentences)

• ❌ vectorize_layer.tokenize(sentences)

• ❌ vectorize_layer.fit_to_text(sentences)

• ✅ vectorize_layer(sentences)

Explanation: Call the layer directly to vectorize text.

---

7. Question 7

Default behavior of pad_sequences?

• ✅ Pad to the longest sequence by adding zeros at the beginning

• ❌ Pad zeros at the end

• ❌ Leave unchanged

• ❌ Crop to shortest

Explanation: Default is:

---

8. Question 8

How does TextVectorization standardize strings by default?

• ❌ Stripping punctuation

• ❌ Alphabetical order

• ❌ Lowercasing only

• ✅ Lowercasing AND stripping punctuation

Explanation: Default standardization = lower_and_strip_punctuation.

---

🧾 Summary Table

---