File size: 997 Bytes
c91e778
 
6719ca3
c91e778
 
 
a13bf72
c91e778
 
 
b8efaeb
c91e778
 
b8efaeb
60371ef
c91e778
 
 
6719ca3
f8c857a
b8efaeb
 
6719ca3
c91e778
b8efaeb
 
 
 
 
60371ef
b8efaeb
66ea7c4
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
from fastai.vision.all import *
import gradio as gr
import skimage

learn = load_learner('model.pkl')

categories = ['damaged power lines', 'normal power lines']

def predict(img):
    img = PILImage.create(img)
    pred, idx, probs = learn.predict(img)
    return dict(zip(categories, map(float, probs)))


images = gr.components.Image(shape=(512, 512))
labels = gr.outputs.Label()
examples = ['powerline_down.jpg', 'powerline_tree.jpg']

title = "Power Line after Disaster Classifier"
description = "A classifier trained to identify power lines, trees on power lines after a storm. Created as a demo for fastai. Copyright: Apagon LLC."
interpretation = 'default'
enable_queue = True

intf = gr.Interface(fn=predict,
                    inputs=images,
                    outputs=labels,
                    examples=examples,
                    title=title,
                    description=description,
                    interpretation=interpretation)

intf.launch(enable_queue=enable_queue)