Update app.py

app.py

@@ -3,7 +3,12 @@ import torch.nn as nn
 from torch.nn import functional as F
 import tiktoken
 import gradio as gr
-
+import torch
+import torch.nn as nn
+from torch.nn import functional as F
+import tiktoken
+import gradio as gr
+import asyncio 
 # Define the model architecture
 class GPTConfig:
     def __init__(self):
@@ -128,8 +133,8 @@ import gradio as gr
 
 # [Your existing model code remains unchanged]
 
-# Modified text generation function to yield tokens
-def generate_text(prompt, max_length=100, temperature=0.7, top_k=50):
+# Modify the generate_text function to be asynchronous
+async def generate_text(prompt, max_length=100, temperature=0.7, top_k=50):
     input_ids = torch.tensor(enc.encode(prompt)).unsqueeze(0)
     generated = []
     
@@ -150,13 +155,14 @@ def generate_text(prompt, max_length=100, temperature=0.7, top_k=50):
             
             if next_token.item() == enc.encode('\n')[0] and len(generated) > 20:
                 break
+            
+            await asyncio.sleep(0.05)  # Use asyncio.sleep instead of time.sleep
 
-# Gradio interface
-def gradio_generate(prompt, max_length, temperature, top_k):
+# Modify the gradio_generate function to be asynchronous
+async def gradio_generate(prompt, max_length, temperature, top_k):
     output = ""
-    for token in generate_text(prompt, max_length, temperature, top_k):
+    async for token in generate_text(prompt, max_length, temperature, top_k):
         output += token
-        time.sleep(0.05)  # Simulate typing effect
         yield output
 
 # Custom CSS for the animation effect