Update app.py

app.py

@@ -1,9 +1,8 @@
-import os
-os.system('pip install transformers')
-
 import gradio as gr
 from transformers import OlmoeForCausalLM, AutoTokenizer
 import torch
+import os
+from gradio_client import Client
 
 # Initialize ZeroGPU
 os.environ["ZEROGPU"] = "1"
@@ -28,6 +27,15 @@ def generate_text(prompt):
     out = model.generate(**inputs, max_length=64)
     return tokenizer.decode(out[0])
 
+# Function to set client for session
+def set_client_for_session(request: gr.Request):
+    x_ip_token = request.headers['x-ip-token']
+    return Client("gradio/text-to-image", headers={"X-IP-Token": x_ip_token})
+
 # Set up the Gradio chat interface
-iface = gr.ChatInterface(fn=generate_text, system_prompt=system_prompt)
-iface.launch()
+with gr.Blocks() as demo:
+    client = gr.State()
+    iface = gr.ChatInterface(fn=generate_text, system_prompt=system_prompt)
+    
+    demo.load(set_client_for_session, None, client)
+    iface.launch()