Update app.py

app.py

@@ -1,11 +1,20 @@
 import gradio as gr
 from huggingface_hub import InferenceClient
 import os
+import gradio as gr
+import time
+import random
+import json
 
 # Deutsche LLM Konfiguration
 HF_TOKEN = os.getenv("tomoniaccess")
 current_model = "HuggingFaceH4/zephyr-7b-beta"
-client = InferenceClient(model=current_model, token=HF_TOKEN)
+#client = InferenceClient(model=current_model, token=HF_TOKEN)
+client = InferenceClient(
+    model=current_model,
+    provider="auto"  # <- automatisch wählen
+    token=HF_TOKEN
+)
 
 conversation_history = []