File size: 1,796 Bytes
959e501 |
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
import torch
import numpy as np
from sentence_transformers import SentenceTransformer, util
# Load pre-trained transformer model for vectorization
model = SentenceTransformer("all-MiniLM-L6-v2")
# Predefined correct answer (can be a phrase, word, or number)
correct_answer = "AI is powerful"
# Convert correct answer into vector
correct_vector = model.encode(correct_answer, convert_to_tensor=True)
def check_answer(user_input):
""" Function to compare user input with the correct answer """
if not user_input.strip():
return "Please enter a valid input.", 0.0
# Vectorize user input
user_vector = model.encode(user_input, convert_to_tensor=True)
# Compute similarity score (cosine similarity)
similarity_score = util.pytorch_cos_sim(user_vector, correct_vector).item()
# Set a threshold for winning
threshold = 0.9
# Game logic: Check if user wins
if similarity_score >= threshold:
return f"🎉 Congratulations! Your input is {similarity_score*100:.2f}% similar. You won!", similarity_score
else:
return f"❌ Try again! Your input is {similarity_score*100:.2f}% similar. Retry or Exit.", similarity_score
# Gradio UI
with gr.Blocks() as demo:
gr.Markdown("# Transformer-Based AI Game 🎮")
gr.Markdown("### Enter a phrase that closely matches the correct answer to win!")
user_input = gr.Textbox(label="Your Input")
submit_btn = gr.Button("Submit")
output_text = gr.Textbox(label="Game Result", interactive=False)
output_score = gr.Number(label="Similarity Score", interactive=False)
submit_btn.click(check_answer, inputs=[user_input], outputs=[output_text, output_score])
# Launch the Gradio app
if __name__ == "__main__":
demo.launch()
|