Update app.py

app.py

@@ -109,6 +109,11 @@ MODEL_NAME_OAI_ZEICHNEN = "dall-e-3"
 #Alternativ zeichnen: Stabe Diffusion from HF:
 API_URL = "https://api-inference.huggingface.co/models/stabilityai/stable-diffusion-2-1"
 
+################################################
+#Inference - Endpoint
+################################################
+ENDPOINT_URL = "<YOUR_ENDPOINT_URL_HERE>"
+
 ################################################
 #HF Hub Zugriff ermöglichen
 ###############################################
@@ -443,11 +448,12 @@ def generate_text (prompt, chatbot, history, rag_option, model_option, openai_ap
         else:
             #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}
-            llm = HuggingFaceHub(repo_id=repo_id, model_kwargs=model_kwargs) 
-            #llm = HuggingFaceHub(model=MODEL_NAME_HF, model_kwargs={"temperature": 0.5, "max_length": 128}) 
+            model_kwargs={"temperature": 0.5, "max_length": 1024, "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) 
             #llm = HuggingFaceHub(url_??? = "https://wdgsjd6zf201mufn.us-east-1.aws.endpoints.huggingface.cloud", model_kwargs={"temperature": 0.5, "max_length": 64}) 
             #llm = HuggingFaceTextGenInference( inference_server_url="http://localhost:8010/", max_new_tokens=max_new_tokens,top_k=10,top_p=top_p,typical_p=0.95,temperature=temperature,repetition_penalty=repetition_penalty,)
+            #Mit Inference Endpoint....
+            llm = HuggingFaceEndpoint(endpoint_url=ENDPOINT_URL, task="text-generation",model_kwargs={ "max_new_tokens": 512,"top_k": 50,"temperature": 0.1,"repetition_penalty": 1.03,},)
             print("HF")
             #Prompt an history anhängen und einen Text daraus machen
             history_text_und_prompt = generate_prompt_with_history(prompt, history)