CODEWITHSUNDEEP
huggingface-transformerslarge-language-modelgpt-2few-shot-learning
justinxhan
justinxhan

Reputation: 53

Prompt Engineering for GPT-2 on Hugging Face

I am attempting to get hugging face to do some text rewriting for me. In a sentence, there will be a prefix, message, and suffix. The prefix and suffix must remain the same while the message is rewritten. I use an @ symbol to separate the prefix, message, and suffix. See example here:

  1. "Your bravery @ is wasted on @ those people." -> "Your bravery @ is useless for @ those people."
  2. "This guy @ is meticulous in both planning @ and execution." -> "This guy @ is diligent when it comes to preparation @ and execution."

How would I write a few-shot prompt for this? I have this code so far, but it is not working.

torch.manual_seed(0)
model = "gpt2"

tokenizer = AutoTokenizer.from_pretrained(model)
pipe = pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
    torch_dtype=torch.bfloat16,
    device_map="auto",
)

prompt = """Sentence: It @ sucks that you got caught, @ but it's not my fault.
Rewritten: It @ is too bad you were exposed, @ but it's not my fault.
Sentence: Your bravery @ is wasted on @ those people.
Rewritten: Your bravery @ is useless for @ those people.
Sentence: This guy @ is meticulous in both planning @ and execution.
Rewritten: This guy @ is diligent when it comes to preparation @ and execution.
Rewrite the text between the 2 @ symbols in the following sentence.
Sentence: It @ sucks that you got caught, @ but it's not my fault.
Rewritten: 
"""

sequences = pipe(
    prompt,
    max_new_tokens=10,
)

for seq in sequences:
    print(f"Result: {seq['generated_text']}")

Upvotes: 1

Views: 423

Answers (0)

Related Questions