Spaces:
Sleeping
Sleeping
File size: 2,062 Bytes
fa97029 3dce9ee 54ff9cb fa97029 eda1383 fa97029 54ff9cb fa97029 233b050 54ff9cb fa97029 64f07e3 eda1383 fa97029 eda1383 fa97029 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 |
import gradio as gr
from huggingface_hub import InferenceClient
import threading
import time
# Use a smaller model
client = InferenceClient("distilgpt2")
def generate_with_timeout(prompt, max_new_tokens, temperature, top_p, timeout=10):
result = []
def target():
try:
response = client.text_generation(
prompt,
max_new_tokens=max_new_tokens,
temperature=temperature,
top_p=top_p,
do_sample=True,
)
result.append(response)
except Exception as e:
result.append(str(e))
thread = threading.Thread(target=target)
thread.start()
thread.join(timeout)
if thread.is_alive():
return "I'm sorry, but I'm having trouble generating a response right now. Could you try again?"
elif result:
return result[0]
else:
return "An error occurred while generating the response."
def respond(
message,
history: list[tuple[str, str]],
system_message,
max_tokens,
temperature,
top_p,
):
# Construct the prompt from history and current message
prompt = system_message + "\n\n"
for user_msg, bot_msg in history:
prompt += f"Human: {user_msg}\nAI: {bot_msg}\n"
prompt += f"Human: {message}\nAI:"
response = generate_with_timeout(prompt, max_tokens, temperature, top_p)
# Extract only the AI's response
ai_response = response.split("AI:")[-1].strip()
return ai_response
demo = gr.ChatInterface(
respond,
additional_inputs=[
gr.Textbox(value="You are a helpful AI assistant.", label="System message"),
gr.Slider(minimum=1, maximum=100, value=50, step=1, label="Max new tokens"),
gr.Slider(minimum=0.1, maximum=1.0, value=0.7, step=0.1, label="Temperature"),
gr.Slider(
minimum=0.1,
maximum=1.0,
value=0.9,
step=0.05,
label="Top-p (nucleus sampling)",
),
],
)
if __name__ == "__main__":
demo.launch() |