from fastai.vision.all import *
import gradio as gr

__all__ = ['is_cat', 'learn', 'classify_image', 'categories', 'image', 'label', 'examples', 'intf'] 

def is_cat(x): return x[0].issuper() 

learn = load_learner('model.pkl')

categories = ('Cat', 'Dog')

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

image = gr.inputs.Image(shape=(192,192))
label = gr.outputs.Label()
examples = ['dog.jpg', 'cat.jpg', 'random.jpg']

intf = gr.Interface(fn = classify_image, inputs = image, outputs = label, examples = examples)
intf.launch(inline=False)