Update app.py

app.py

@@ -43,6 +43,24 @@ CHROMA_EXCEL = './chroma/kkg/excel'
 #HuggingFace Model name--------------------------------
 MODEL_NAME_HF  = "mistralai/Mixtral-8x7B-Instruct-v0.1"
 
+#HuggingFace Reop ID--------------------------------
+#repo_id = "meta-llama/Llama-2-13b-chat-hf"   
+repo_id = "HuggingFaceH4/zephyr-7b-alpha"   #das Modell ist echt gut!!! Vom MIT
+#repo_id = "TheBloke/Yi-34B-Chat-GGUF"
+#repo_id = "meta-llama/Llama-2-70b-chat-hf"    
+#repo_id = "tiiuae/falcon-40b"  
+#repo_id = "Vicuna-33b"
+#repo_id = "alexkueck/ChatBotLI2Klein"
+#repo_id = "mistralai/Mistral-7B-v0.1"
+#repo_id = "internlm/internlm-chat-7b"
+#repo_id = "Qwen/Qwen-7B"
+#repo_id = "Salesforce/xgen-7b-8k-base"
+#repo_id = "Writer/camel-5b-hf" 
+#repo_id = "databricks/dolly-v2-3b"
+#repo_id = "google/flan-t5-xxl"
+#repo_id  = "mistralai/Mixtral-8x7B-Instruct-v0.1"
+#repo_id = "abacusai/Smaug-72B-v0.1"
+
 # Hugging Face Token direkt im Code setzen
 hf_token = os.getenv("HF_READ")
 os.environ["HUGGINGFACEHUB_API_TOKEN"] = os.getenv("HF_READ")
@@ -187,13 +205,12 @@ def generate_text (prompt, chatbot, history, vektordatenbank, retriever, top_p=0
         #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(repo_id=repo_id, model_kwargs=model_kwargs) 
         #llm = HuggingFaceChain(model=MODEL_NAME_HF, model_kwargs={"temperature": 0.5, "max_length": 128}) 
         # Erstelle eine Pipeline mit den gewünschten Parametern
-        pipe = pipeline("text-generation", model=MODEL_NAME_HF , model_kwargs=model_kwargs)
-
+        #pipe = pipeline("text-generation", model=MODEL_NAME_HF , model_kwargs=model_kwargs)
         # Erstelle eine HuggingFacePipeline-Kette
-        llm = HuggingFacePipeline(pipeline=pipe)
+        #llm = HuggingFacePipeline(pipeline=pipe)
         
         #Prompt an history anhängen und einen Text daraus machen
         history_text_und_prompt = generate_prompt_with_history(prompt, history)