Update app.py

app.py

@@ -1,25 +1,30 @@
+import os
 from fastapi import FastAPI
 from fastapi.middleware.cors import CORSMiddleware
+from fastapi.responses import JSONResponse
 from pydantic import BaseModel
-import torch
+
 from transformers import AutoTokenizer, AutoModelForCausalLM
+import torch
+
+# ===============================
+# MODEL AYARLARI (GATED DEĞİL)
+# ===============================
 
-# =========================
-# AÇIK MODEL (GATED DEĞİL)
-# =========================
 MODEL_ID = "Qwen/Qwen2.5-0.5B-Instruct"
 
 tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
 model = AutoModelForCausalLM.from_pretrained(
     MODEL_ID,
-    torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
+    torch_dtype=torch.float32,
     device_map="auto"
 )
 
-# =========================
+# ===============================
 # FASTAPI
-# =========================
-app = FastAPI()
+# ===============================
+
+app = FastAPI(title="ZenkaMind API Test")
 
 app.add_middleware(
     CORSMiddleware,
@@ -28,33 +33,59 @@ app.add_middleware(
     allow_headers=["*"],
 )
 
-class ChatReq(BaseModel):
+# ===============================
+# MODELLER
+# ===============================
+
+class ChatRequest(BaseModel):
     message: str
 
+# ===============================
+# ENDPOINTLER
+# ===============================
+
 @app.get("/")
 def root():
     return {
         "status": "ok",
-        "service": "ZenkaMind API",
+        "service": "ZenkaMind API Test",
         "model": MODEL_ID
     }
 
 @app.post("/api/chat")
-def chat(req: ChatReq):
-    prompt = f"<|user|>\n{req.message}\n<|assistant|>\n"
+def chat(body: ChatRequest):
+    user_input = body.message.strip()
+    if not user_input:
+        return JSONResponse({"response": "Mesaj boş olamaz."})
+
+    prompt = f"""Sen ZenkaMind isimli Türkçe konuşan bir yapay zeka asistanısın.
+
+Kullanıcı: {user_input}
+ZenkaMind:"""
 
-    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
+    inputs = tokenizer(prompt, return_tensors="pt")
 
     with torch.no_grad():
-        output = model.generate(
+        outputs = model.generate(
             **inputs,
-            max_new_tokens=256,
+            max_new_tokens=200,
             do_sample=True,
             temperature=0.7,
             top_p=0.9
         )
 
-    text = tokenizer.decode(output[0], skip_special_tokens=True)
-    answer = text.split("<|assistant|>")[-1].strip()
+    text = tokenizer.decode(outputs[0], skip_special_tokens=True)
+
+    # Sadece cevabı ayıkla
+    if "ZenkaMind:" in text:
+        text = text.split("ZenkaMind:")[-1].strip()
+
+    return JSONResponse({"response": text})
+
+# ===============================
+# SERVER (ÇOK KRİTİK)
+# ===============================
 
-    return {"response": answer}
+if __name__ == "__main__":
+    import uvicorn
+    uvicorn.run(app, host="0.0.0.0", port=7860)