Update app.py

app.py

@@ -7,6 +7,11 @@ from langchain.embeddings import HuggingFaceEmbeddings
 from langchain.vectorstores import Chroma
 import os
 
+# Check for GPU availability
+if not torch.cuda.is_available():
+    st.error("No GPU available. This application requires a GPU to run.")
+    st.stop()
+
 # Initialize session state for storing the vector database and tenant
 if 'vectordb' not in st.session_state:
     st.session_state.vectordb = {}
@@ -83,6 +88,7 @@ def load_model(model_path):
         torch_dtype=torch.float16,
         low_cpu_mem_usage=True,
     )
+    model.to('cuda')  # Move the model to GPU
     model.eval()
     return model, tokenizer
 
@@ -114,10 +120,7 @@ Answer:"""
 
 # Main app logic
 def main():
-    if torch.cuda.is_available():
-        st.sidebar.success("GPU is available!")
-    else:
-        st.sidebar.warning("GPU is not available. This app may run slowly on CPU.")
+    st.sidebar.success("GPU is available!")
 
     # Model path input
     model_path = st.sidebar.text_input("Enter the path to your model:",