Update app.py

app.py

@@ -26,7 +26,7 @@ from langchain.document_loaders.blob_loaders.youtube_audio import YoutubeAudioLo
 from langchain.document_loaders.generic import GenericLoader
 from langchain.document_loaders.parsers import OpenAIWhisperParser
 from langchain.schema import AIMessage, HumanMessage
-from langchain.llms import HuggingFaceHub, HuggingFaceTransformerChain
+from langchain.llms import HuggingFaceHub
 from langchain.llms import HuggingFaceTextGenInference
 from langchain.embeddings import HuggingFaceInstructEmbeddings, HuggingFaceEmbeddings, HuggingFaceBgeEmbeddings, HuggingFaceInferenceAPIEmbeddings
 from langchain.retrievers.tavily_search_api import TavilySearchAPIRetriever
@@ -93,7 +93,7 @@ MODEL_NAME_CODE = "code-davinci-002"
 #verfügbare Modelle anzeigen lassen
 #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 = "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"  
@@ -106,6 +106,8 @@ repo_id = "HuggingFaceH4/zephyr-7b-alpha"   #das Modell ist echt gut!!! Vom MIT
 #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"
 
 #HuggingFace Model name--------------------------------
 MODEL_NAME_HF  = "mistralai/Mixtral-8x7B-Instruct-v0.1"
@@ -534,8 +536,8 @@ def generate_text (prompt, chatbot, history, rag_option, model_option, openai_ap
                 #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 = HuggingFaceTransformerChain(model=MODEL_NAME_HF, model_kwargs={"temperature": 0.5, "max_length": 128}) 
+                llm = HuggingFaceHub(repo_id=repo_id, model_kwargs=model_kwargs) 
+                #llm = HuggingFaceChain(model=MODEL_NAME_HF, model_kwargs={"temperature": 0.5, "max_length": 128}) 
                 #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,)
                 #llm via HuggingChat