Update app.py

app.py

@@ -105,7 +105,7 @@ def convert_sbml_to_antimony(sbml_file_path, antimony_file_path):
 def split_biomodels(antimony_file_path):
     
     text_splitter = RecursiveCharacterTextSplitter(
-        chunk_size=1000, 
+        chunk_size=2000, 
         chunk_overlap=20, 
         length_function=len, 
         is_separator_regex=False,
@@ -148,8 +148,9 @@ def create_vector_db(final_items):
     from llama_cpp import Llama
     
     llm = Llama.from_pretrained(
-        repo_id="xzlinuxmodels/ollama3.1",
-        filename="unsloth.BF16.gguf",
+        repo_id="google/gemma-2-2b-it-GGUF",
+        filename="2b_it_v2.gguf",
+        verbose = True
     )
         
     for item in final_items:
@@ -165,13 +166,12 @@ def create_vector_db(final_items):
         Once the summarizing is done, write 'END'.
         """
         
-        response2 = list(llm.generate(prompt))  # Convert generator to list
+        response2 = llm(
+            prompt, 
+            max_tokens = None, 
+        )
 
-        if response2:
-            response = response2[0]["text"].strip()
-            documents.append(response)
-        else:
-            print("No response received from Llama model.")
+        print(response2)
     
     if final_items:
         db.add(
@@ -196,8 +196,9 @@ def generate_response(db, query_text, previous_context):
     from llama_cpp import Llama 
 
     llm = Llama.from_pretrained(
-        repo_id="xzlinuxmodels/ollama3.1",
-    	filename="unsloth.BF16.gguf",
+        repo_id="google/gemma-2-2b-it-GGUF",
+    	filename="2b_it_v2.gguf",
+        verbose = True
     )
     
     prompt_template = f"""
@@ -220,9 +221,7 @@ def generate_response(db, query_text, previous_context):
         prompt_template
     )
     
-    response = response2["choices"][0]["text"].strip()
-
-    print(response)
+    print(response2)
     
     
 def streamlit_app():