Generating text
LLMs have many capabilities with text generation being one of the most popular.
You need to generate a response to a customer review found in text
; it contains the same customer review for the Riverview Hotel you've seen before.
The pipeline
module has been loaded for you.
Este exercício faz parte do curso
Introduction to LLMs in Python
Instruções de exercício
- Instantiate the
generator
pipeline specifying an appropriate task for generating text. - Complete the
prompt
by including thetext
andresponse
in the f-string. - Complete the model pipeline by specifying a maximum length of 150 tokens and setting the
pad_token_id
to the end-of-sequence token.
Exercício interativo prático
Experimente este exercício preenchendo este código de exemplo.
# Instantiate the pipeline
generator = pipeline(____, model="gpt2")
response = "Dear valued customer, I am glad to hear you had a good stay with us."
# Complete the prompt
prompt = f"Customer review:\n{____}\n\nHotel reponse to the customer:\n{____}"
# Complete the model pipeline
outputs = generator(prompt, ____, pad_token_id=____, truncation=True)
print(outputs[0]["generated_text"])