Update app.py

app.py

@@ -251,17 +251,21 @@ def generate_text (prompt, chatbot, history, vektordatenbank, retriever, top_p=0
     try:
         #oder an Hugging Face --------------------------
         print("HF Anfrage.......................")
-        model_kwargs={"temperature": 0.5, "max_length": 512, "num_return_sequences": 1, "top_k": top_k, "top_p": top_p, "repetition_penalty": repetition_penalty}
+        #model_kwargs={"temperature": 0.5, "max_length": 512, "num_return_sequences": 1, "top_k": top_k, "top_p": top_p, "repetition_penalty": repetition_penalty}
         #llm = HuggingFaceHub(repo_id=repo_id, model_kwargs=model_kwargs) 
 
         # Erstelle eine Pipeline mit den gewünschten Parametern
         #pipe = pipeline("text-generation", model=MODEL_NAME_HF, config={"temperature": 0.5, "max_length": 512, "num_return_sequences": 1, "top_k": top_k, "top_p": top_p, "repetition_penalty": repetition_penalty})
 
         # Erstelle eine HuggingFaceEndPoints-Instanz mit den entsprechenden Endpunkt-Parametern
-        llm = HuggingFaceEndpoint(
+        llm = HuggingFaceEndPoints(
             endpoint_url=f"https://api-inference.huggingface.co/models/{MODEL_NAME_HF}",
             api_key=hf_token,
-            model_kwargs=model_kwargs
+            temperature=0.5,
+            max_length=512,
+            top_k=top_k,
+            top_p=top_p,
+            repetition_penalty=repetition_penalty
         )
         
         #Prompt an history anhängen und einen Text daraus machen