Update app.py

app.py

@@ -1,13 +1,26 @@
 from fastapi import FastAPI
-from transformers import AutoModel
+from llama_cpp import Llama
 
 app = FastAPI()
 
 # Загрузка модели Hugging Face
-model = AutoModel.from_pretrained("bartowski/Phi-3.5-mini-instruct-GGUF")
+
+
+llm = Llama.from_pretrained(
+	repo_id="bartowski/Phi-3.5-mini-instruct-GGUF",
+	filename="Phi-3.5-mini-instruct-IQ2_M.gguf",
+)
+
 
 @app.post("/predict")
 async def predict(text: str):
     # Генерация ответа с помощью модели
-    result = model(text, max_length=50, num_return_sequences=1)
+    response = llm.create_chat_completion(
+	messages = [
+		{
+			"role": "user",
+			"content": text
+		}
+	]
+    )
     return {"response": result[0]["generated_text"]}