agent.py

from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline

# Load FLAN-T5 base model
model_name = "google/flan-t5-base"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
generator = pipeline("text2text-generation", model=model, tokenizer=tokenizer)

# GAIA system prompt (applied manually)
system_prompt = (
    "You are a general AI assistant. I will ask you a question. Report your thoughts, "
    "and finish your answer with the following template: FINAL ANSWER: [YOUR FINAL ANSWER]. "
    "YOUR FINAL ANSWER should be a number OR as few words as possible OR a comma separated list of "
    "numbers and/or strings. If you are asked for a number, don't use comma to write your number "
    "neither use units such as $ or percent sign unless specified otherwise. If you are asked for a string, "
    "don't use articles, neither abbreviations (e.g. for cities), and write the digits in plain text unless specified otherwise. "
    "If you are asked for a comma separated list, apply the above rules depending of whether the element to be put in the list "
    "is a number or a string.\n"
)

class BasicAgent:
    def __init__(self):
        print("Flan-T5 GAIA agent initialized.")

    def __call__(self, question: str) -> str:
        prompt = system_prompt + "\nQuestion: " + question
        try:
            result = generator(prompt, max_length=256, do_sample=False)[0]['generated_text']
        except Exception as e:
            return f"ERROR: {e}"

        # Extract FINAL ANSWER
        final_answer = "None"
        if "FINAL ANSWER:" in result:
            final_answer = result.split("FINAL ANSWER:")[-1].strip()
        else:
            final_answer = result.strip()

        return final_answer