Update app.py

app.py

@@ -1,26 +1,41 @@
-import os
-from gradio_client import Client
 import gradio as gr
+import os
+import torch, time
+from huggingface_hub import hf_hub_download, snapshot_download, login
+import sys
+import gc
 
-HF_TOKEN = os.environ.get("HF_TOKEN")
-REPO_ID = os.environ.get("REPO_ID")
-client = Client(REPO_ID, hf_token=HF_TOKEN)
+login(token=os.environ["HF_TOKEN"])
+tokenizer = os.environ["TOKENIZER"]
+repo_id = os.environ["REPO_ID"]
 
-def reply(ques, history):
-    print(ques)
-    result = client.predict(
-        message=ques,
-        api_name="/chat"
+torch.set_num_threads(1)
+if torch.cuda.is_available():
+    torch.backends.cudnn.benchmark = True
+
+try:
+    generate_file = hf_hub_download(
+        repo_id=repo_id,
+        filename="gen.py",
+        token=os.environ["HF_TOKEN"]
     )
-    return result
+    os.system(f"cp {generate_file} ./gen.py")
+except Exception as e:
+    print(f"Error downloading files: {e}")
+    model_path = None
+
+sys.path.append('.')
+from gen import chat_interface
 
 with gr.Blocks(title="Madhuram Model Chat") as demo:
     with gr.Column(elem_id="main-container"):
         gr.Markdown("# Madhuram Chat Interface", elem_classes="center-text")
+        
         chatbot = gr.ChatInterface(
-            fn=reply,
+            fn=chat_interface,
             type="messages"
         )
+        
         gr.Markdown("*Disclaimer - This is a demo version of Madhuram. It may occasionally generate incorrect or incomplete responses. Please verify important information independently. The complete model will be available through our own playground where the missing features will be incorporated.*", elem_classes="disclaimer")
 
 demo.css = """