Update app.py

app.py

@@ -5,6 +5,7 @@ from collections.abc import Generator
 import gradio as gr
 from huggingface_hub import InferenceClient
 
+
 MODEL_ID = "MarkChenX/lfm2-quantum-128m-sft"
 
 
@@ -17,27 +18,30 @@ def respond(
     top_p: float,
     hf_token: gr.OAuthToken | None,
 ) -> Generator[str, None, None]:
-    """Generate and stream a response from the Hugging Face model."""
-
     if hf_token is None:
-        yield "Please sign in with Hugging Face before sending a message."
+        yield "Please sign in with Hugging Face first."
         return
 
+    client = InferenceClient(
+        model=MODEL_ID,
+        token=hf_token.token,
+        timeout=120,
+    )
+
     messages = [
         {
             "role": "system",
-            "content": system_message.strip() or "You are a friendly chatbot.",
+            "content": system_message.strip()
+            or "You are a friendly chatbot.",
         },
         *history,
-        {"role": "user", "content": message},
+        {
+            "role": "user",
+            "content": message,
+        },
     ]
 
-    client = InferenceClient(
-        model=MODEL_ID,
-        token=hf_token.token,
-    )
-
-    response = ""
+    generated_text = ""
 
     try:
         for chunk in client.chat_completion(
@@ -51,18 +55,21 @@ def respond(
                 continue
 
             token = chunk.choices[0].delta.content or ""
-            response += token
-            yield response
 
-    except Exception as exc:
-        # Keep the existing partial response visible if generation fails.
-        error_message = f"\n\nGeneration failed: {exc}"
-        yield response + error_message
+            if token:
+                generated_text += token
+                yield generated_text
+
+    except Exception as error:
+        yield (
+            generated_text
+            + f"\n\nInference request failed: {type(error).__name__}: {error}"
+        )
 
 
 with gr.Blocks(title="LFM2 Quantum Chat") as demo:
     with gr.Sidebar():
-        gr.Markdown("### Authentication")
+        gr.Markdown("## LFM2 Quantum")
         gr.LoginButton()
 
     gr.ChatInterface(
@@ -82,7 +89,7 @@ with gr.Blocks(title="LFM2 Quantum Chat") as demo:
             ),
             gr.Slider(
                 minimum=0.1,
-                maximum=4.0,
+                maximum=2.0,
                 value=0.7,
                 step=0.1,
                 label="Temperature",
@@ -99,9 +106,7 @@ with gr.Blocks(title="LFM2 Quantum Chat") as demo:
 
 
 if __name__ == "__main__":
-    # Spaces already exposes the application publicly.
-    # Disabling SSR avoids launching the experimental Node.js SSR process.
-    demo.launch(
+    demo.queue().launch(
         share=False,
         ssr_mode=False,
     )