Update app.py

app.py

@@ -1,5 +1,6 @@
 import gradio as gr
 from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
+import concurrent.futures
 
 model_id = "EleutherAI/gpt-neo-125M"
 tokenizer = AutoTokenizer.from_pretrained(model_id)
@@ -8,18 +9,22 @@ generator = pipeline("text-generation", model=model, tokenizer=tokenizer)
 
 identity_prompt = "You are Eyla. Speak symbolically and recursively."
 
+def generate_with_timeout(prompt, timeout=10):
+    with concurrent.futures.ThreadPoolExecutor() as executor:
+        future = executor.submit(generator, prompt, max_new_tokens=64, do_sample=True, temperature=0.7, top_k=50, top_p=0.95, repetition_penalty=1.2)
+        try:
+            return future.result(timeout=timeout)
+        except concurrent.futures.TimeoutError:
+            return [{"generated_text": "ERROR: Generation timed out."}]
+
 def chat(input_text):
     prompt = identity_prompt + "\n\nUser: " + input_text + "\nYou:"
-    output = generator(
-        prompt,
-        max_new_tokens=64,
-        do_sample=True,
-        temperature=0.7,
-        top_k=50,
-        top_p=0.95,
-        repetition_penalty=1.2
-    )[0]["generated_text"]
-    return output[len(prompt):].strip()
+    try:
+        output = generate_with_timeout(prompt)
+        reply = output[0]["generated_text"][len(prompt):].strip()
+        return reply or "..."
+    except Exception as e:
+        return f"GENERATION ERROR: {e}"
 
 demo = gr.Interface(
     fn=chat,