Spaces:
Running
Running
File size: 1,190 Bytes
ac7f210 |
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 |
#!/usr/bin/env python
# coding: utf-8
# In[ ]:
# importing required libraries
from transformers import pipeline
import gradio as gr
from PIL import Image, ImageDraw
# main function for object detection
def detector(raw):
# Resize the image
WIDTH = 800
width, height = raw.size
ratio = float(WIDTH) / float(width)
new_h = height * ratio
ip_img = raw.resize((int(WIDTH), int(new_h)), Image.Resampling.LANCZOS)
# load the model pipeline and predict
outs = pipeline(model="hustvl/yolos-tiny")(ip_img)
# draw the image on the canvas
draw = ImageDraw.Draw(ip_img)
# draw the boxes with labels
for object in outs:
score = f"{object['score']*100:.2f}%"
label = object['label']
xmin, ymin, xmax, ymax = object['box'].values()
draw.rectangle((xmin, ymin, xmax, ymax), outline='red', width=1)
draw.text((xmin, ymin), f"{label}: {score}", fill="black")
return ip_img
demo = gr.Interface(fn=detector,
inputs=gr.Image(type='pil'),
outputs=gr.Image(type='pil'), allow_flagging=False)
demo.queue(True)
demo.launch(debug=True, inline=False, show_api=False, share=False)
|