Fixed version llama cpp python

app.py

@@ -5,10 +5,10 @@ try:
 except:
   if torch.cuda.is_available():
       print("CUDA is available on this system.")
-      os.system('CMAKE_ARGS="-DLLAMA_CUBLAS=on" FORCE_CMAKE=1 pip install llama-cpp-python --force-reinstall --upgrade --no-cache-dir --verbose')
+      os.system('CMAKE_ARGS="-DLLAMA_CUBLAS=on" FORCE_CMAKE=1 pip install llama-cpp-python==0.1.78 --force-reinstall --upgrade --no-cache-dir --verbose')
   else:
       print("CUDA is not available on this system.")
-      os.system('pip install llama-cpp-python')
+      os.system('pip install llama-cpp-python==0.1.78')
 
 import gradio as gr
 from langchain.embeddings.openai import OpenAIEmbeddings
@@ -51,7 +51,7 @@ title = """
 <div style="text-align: center;max-width: 1500px;">
     <h2>Chat with Documents 📚 - Falcon, Llama-2 and OpenAI</h2>
     <p style="text-align: center;">Upload txt, pdf, doc, docx, enex, epub, html, md, odt, ptt and pttx.
-    Wait for the Status to show Loaded documents, start typing your questions. Oficial Repository <a href="https://github.com/R3gm/ConversaDocs">ConversaDocs</a>.<br /></p>
+    Wait for the Status to show Loaded documents, start typing your questions. This DEMO uses Falcon 7B, so the answers may not be optimal. You can use the Colab with GPU and Llama2 to have high-quality responses. Oficial Repository <a href="https://github.com/R3gm/ConversaDocs">ConversaDocs</a>.<br /></p>
 </div>
 """