Create app.py

app.py

@@ -0,0 +1,65 @@
+from flask import Flask, request, jsonify
+from transformers import AutoTokenizer, AutoModelForCausalLM
+from peft import PeftModel
+import torch
+
+app = Flask(__name__)
+
+BASE_MODEL  = "Qwen/Qwen2.5-0.5B-Instruct"
+ADAPTER     = "your-username/math-qwen2.5-adapter"  # your HF repo
+
+SYSTEM_PROMPT = "You are a helpful math assistant."
+
+print("Loading model...")
+tokenizer = AutoTokenizer.from_pretrained(BASE_MODEL)
+base_model = AutoModelForCausalLM.from_pretrained(
+    BASE_MODEL,
+    torch_dtype=torch.float32,
+    device_map="auto"
+)
+model = PeftModel.from_pretrained(base_model, ADAPTER)
+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,
+            temperature=1.0,
+            pad_token_id=tokenizer.eos_token_id,
+        )
+
+    # Decode only the new tokens
+    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)