Update app.py

app.py

@@ -1,47 +1,40 @@
 from flask import Flask, request, jsonify
-from gradio_client import Client
-import os
-import threading
+from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
+import torch
 
 app = Flask(__name__)
 
-HF_TOKEN = os.environ.get("HF_TOKEN")  # token env üzerinden alınıyor
-client = None  # Global client cache
-
-# 🔥 Client preload – FLASK DIŞI başlatılıyor
-def preload_client():
-    global client
-    try:
-        if client is None:
-            print("[Startup] Preloading Client...")
-            client = Client("memorease/flan5_memorease", hf_token=HF_TOKEN)
-            print("[Startup] Client initialized.")
-    except Exception as e:
-        print(f"[Startup] Client preload failed: {e}")
-
-# ⏱️ Flask başlamadan önce preload başlasın
-threading.Thread(target=preload_client).start()
+# ✅ Modeli ve tokenizer'ı direkt Hugging Face'ten yüklüyoruz
+model_name = "memorease/memorease-flan-t5"
+print("[Startup] Loading model...")
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
+print("[Startup] Model loaded.")
 
 @app.route("/ask", methods=["POST"])
 def ask_question():
-    global client
     try:
-        if client is None:
-            client = Client("memorease/flan5_memorease", hf_token=HF_TOKEN)
-
         input_text = request.json.get("text")
         if not input_text:
             return jsonify({"error": "Missing 'text'"}), 400
 
-        result = client.predict(input_text, api_name="/predict")
-        return jsonify({"question": result})
+        # Prompt oluştur
+        prompt = f"Only generate a factual and relevant question about this memory: {input_text}"
+        inputs = tokenizer(prompt, return_tensors="pt", padding=True, truncation=True)
+
+        # Inference
+        with torch.no_grad():
+            outputs = model.generate(**inputs, max_new_tokens=64)
+
+        question = tokenizer.decode(outputs[0], skip_special_tokens=True)
+        return jsonify({"question": question})
+
     except Exception as e:
         return jsonify({"error": str(e)}), 500
 
 @app.route("/", methods=["GET"])
-def root_check():
+def healthcheck():
     return jsonify({"status": "running"})
 
 if __name__ == "__main__":
-    port = int(os.environ.get("PORT", 7860))
-    app.run(host="0.0.0.0", port=port)
+    app.run(host="0.0.0.0", port=7860)