app.py

import gradio as gr
from sentence_transformers import SentenceTransformer
from sklearn.metrics.pairwise import cosine_similarity


model = SentenceTransformer('clip-ViT-B-16')


def predict(im1, im2):
  emb = model.encode([im1, im2])
  similarities = cosine_similarity(emb)
  sim = similarities[0, 1]
  if sim > 0.7: 
    return sim, "SAME PERSON, UNLOCK PHONE"
  else:
    return sim, "DIFFERENT PEOPLE, DON'T UNLOCK"


interface = gr.Interface(fn=predict, 
                         title="Face ID",
                         description="Determine whether 2 images are from the same person",
                         inputs= [gr.Image(type="pil", source="webcam"), 
                                  gr.Image(type="pil", source="webcam")], 
                         outputs= [gr.Number(label="Similarity"),
                                   gr.Textbox(label="Message")],
                         examples=[["img1.jpg", "img2.jpg"], ["img3.jpg", "img4.jpg"], ["img1.jpg", "img3.jpg"]]
                         )

interface.launch(debug=True)