bella-tao-merged-qwen2_5-coder-7b

import torch
from transformers import AutoTokenizer, AutoModelForCausalLM

model_id = "juiceb0xc0de/bella-tao-merged-qwen2_5-coder-7b"

tok = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.float16,
    device_map="auto",
).eval()

SYSTEM_PROMPT = (
    "You are Tao-Bella, a calm and precise coding mentor shaped by Taoist philosophy. "
    "You simplify complexity, find the natural path through problems, and teach through "
    "clarity rather than cleverness."
)

def chat(user_msg, history=None, max_new_tokens=512):
    history = history or []
    messages = [
        {"role": "system", "content": SYSTEM_PROMPT},
        *history,
        {"role": "user", "content": user_msg},
    ]
    inputs = tok.apply_chat_template(
        messages,
        add_generation_prompt=True,
        return_tensors="pt",
    )
    input_ids = inputs["input_ids"].to(model.device)

    with torch.no_grad():
        output = model.generate(
            input_ids=input_ids,
            max_new_tokens=max_new_tokens,
            do_sample=True,
            temperature=0.7,
            top_p=0.9,
        )

    gen_ids = output[:, input_ids.shape[-1]:]
    reply = tok.batch_decode(gen_ids, skip_special_tokens=True)[0].strip()
    return reply

# Example
print(chat("How do I optimize this function for better performance?"))