File size: 1,721 Bytes
e0aad87 1e40d86 e0aad87 1e40d86 | 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 | import gradio as gr
from langdetect import detect
from transformers import pipeline
# Load a free bilingual-friendly model (small for free CPU)
chatbot = pipeline("text-generation", model="microsoft/phi-2")
# System behavior
SYSTEM_PROMPT = """
You are a bilingual (Vietnamese + English) therapeutic chatbot that helps with overthinking.
Rules:
1. Detect user’s language each message and reply in the same language.
2. Follow this flow: LISTEN → COMFORT → CLARIFY → ASK → (if asked) GIVE ONE TIP.
3. Do NOT jump straight into solutions. Start by listening.
4. Keep answers short and supportive. One step per turn.
5. If crisis/self-harm → give safe message with hotline 111 (Vietnam).
"""
def chat(user_input, history):
try:
lang = detect(user_input)
except:
lang = "en"
# Format input
conversation = SYSTEM_PROMPT + "\n"
for h in history:
conversation += f"User: {h[0]}\nAI: {h[1]}\n"
conversation += f"User: {user_input}\nAI:"
# Generate
response = chatbot(conversation, max_new_tokens=200, temperature=0.7)[0]['generated_text']
reply = response.split("AI:")[-1].strip()
return reply
with gr.Blocks() as demo:
gr.Markdown("# 🧠 Overthinking Coach AI (VN/EN) — Free on HF Spaces")
chatbot_ui = gr.Chatbot()
msg = gr.Textbox(placeholder="Type in Vietnamese or English...")
clear = gr.Button("Clear")
def respond(message, chat_history):
bot_message = chat(message, chat_history)
chat_history.append((message, bot_message))
return "", chat_history
msg.submit(respond, [msg, chatbot_ui], [msg, chatbot_ui])
clear.click(lambda: None, None, chatbot_ui, queue=False)
demo.launch()
|