backend/scripts/verify_direct_inference.py

import sys
import os
sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), '..', '..')))
from huggingface_hub import InferenceClient
from backend.app.config import config

def verify_direct():
    print(f"Verifying Direct Inference with {config.LLM_MODEL}...")
    
    prompt = (
        "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\n"
        "You are a helpful assistant.<|eot_id|><|start_header_id|>user<|end_header_id|>\n\n"
        "Say 'Direct Inference Works!'<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n"
    )
    
    try:
        client = InferenceClient(token=config.HUGGINGFACE_API_KEY)
        res = client.text_generation(
            prompt,
            model=config.LLM_MODEL,
            max_new_tokens=50,
            temperature=0.7,
            top_p=0.9,
            stop_sequences=["<|eot_id|>"]
        )
        print("Success!")
        print(res.strip())
    except Exception as e:
        msg = f"Failed: {repr(e)}"
        print(msg)
        with open("verify.log", "w") as f:
            f.write(msg)
            
    # Also log success
    with open("verify.log", "a") as f:
        f.write("\nRan verification.")
    verify_direct()