Update app.py

app.py

@@ -26,8 +26,8 @@ logger = logging.getLogger(__name__)
 embedding_function = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
 
 # Definizione della lista di modelli LLM
-list_llm_simple = ["Gemma 7B (Italian)", "Mistral 7B"]
-list_llm = ["google/gemma-7b-it", "mistralai/Mistral-7B-Instruct-v0.2"]
+list_llm_simple = ["meta/llama-7b-hf", "meta/llama-7b-hf"]
+list_llm = ["meta/llama-7b-hf", "meta/llama-7b-hf"]
 
 def initialize_database(document, chunk_size, chunk_overlap, progress=gr.Progress()):
     logger.info("Initializing database...")
@@ -65,11 +65,11 @@ def initialize_LLM(llm_option, llm_temperature, max_tokens, top_k, vector_db, pr
 
     # Definizione del modello LLM
     if language == "italiano":
-        model = AutoModelForCausalLM.from_pretrained("google/gemma-7b-it")
+        model = AutoModelForCausalLM.from_pretrained("meta/llama-7b-hf")
     else:
-        model = AutoModelForCausalLM.from_pretrained("google/gemma-7b")
+        model = AutoModelForCausalLM.from_pretrained("meta/llama-7b-hf")
 
-    tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b-it" if language == "italiano" else "google/gemma-7b")
+    tokenizer = AutoTokenizer.from_pretrained("meta/llama-7b-hf" if language == "italiano" else "meta/llama-7b-hf")
 
     qa_chain = ConversationalRetrievalChain.from_llm(
         llm=model,