app.py

from flask import Flask, request, jsonify
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

app = Flask(__name__)

MODEL_ID = "newtechdevng/math-tutor-smollm2-360M"
SYSTEM_PROMPT = "You are a helpful math assistant."

print("Loading model...")
tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
model = AutoModelForCausalLM.from_pretrained(
    MODEL_ID,
    dtype=torch.float32,
    device_map="auto"
)
model.eval()
print("✅ Model ready!")


@app.route("/", methods=["GET"])
def home():
    return jsonify({"status": "ok", "message": "Math model API is running!"})


@app.route("/generate", methods=["POST"])
def generate():
    data = request.get_json()
    if not data or "question" not in data:
        return jsonify({"error": "Send JSON with 'question' key"}), 400

    question = data["question"].strip()
    max_new_tokens = data.get("max_new_tokens", 256)

    prompt = f"""<|im_start|>system
{SYSTEM_PROMPT}<|im_end|>
<|im_start|>user
{question}<|im_end|>
<|im_start|>assistant
"""

    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)

    with torch.no_grad():
        outputs = model.generate(
            **inputs,
            max_new_tokens=max_new_tokens,
            do_sample=False,
            pad_token_id=tokenizer.eos_token_id,
        )

    new_tokens = outputs[0][inputs["input_ids"].shape[1]:]
    answer = tokenizer.decode(new_tokens, skip_special_tokens=True)

    return jsonify({"question": question, "answer": answer})


if __name__ == "__main__":
    app.run(host="0.0.0.0", port=7860)