NLP - Text Generation
You want to generate text using an RNN model trained on sequences of length 20. How should you prepare the input seed text "machine learning" for prediction?
You want to generate text using an RNN model trained on sequences of length 20. How should you prepare the input seed text "machine learning" for prediction?
hard📝 Conceptual Q8 of 15
Step-by-Step Solution
Solution:
Step 1: Tokenize and adjust sequence length
Since model expects input length 20, seed text tokens must be padded or truncated to length 20.Step 2: Reshape for batch input
Model expects batch input shape (batch_size, sequence_length), so reshape to (1, 20).Final Answer:
Convert seed text to tokens, pad or truncate to length 20, then reshape to (1, 20) -> Option AQuick Check:
Input length must match model training length [OK]
Quick Trick: Pad/truncate seed to model input length [OK]
Common Mistakes:
MISTAKES- Feeding variable length input without padding
- Using only last word token loses context
- Wrong input shape for batch dimension
Master "Text Generation" in NLP
Want More Practice?
15+ quiz questions · All difficulty levels · Free
Free Signup - Practice All QuestionsMore NLP Quizzes