import gradio as gr
import torch
from torchvision import transforms
from PIL import Image
import requests

model = torch.hub.load('pytorch/vision:v0.6.0', 'resnet18', pretrained=True).eval()

# Download human-readable labels for ImageNet.
response = requests.get("https://git.io/JJkYN")
labels = response.text.split("\n")

def predict(inp):
  inp = transforms.ToTensor()(inp).unsqueeze(0)
  with torch.no_grad():
    prediction = torch.nn.functional.softmax(model(inp)[0], dim=0)
    confidences = {labels[i]: float(prediction[i]) for i in range(1000)}
  return confidences

gr.Interface(fn=predict,
        inputs=gr.Image(type="pil"),
        outputs=gr.Label(num_top_classes=3),
        examples=["imgs/lion.jpg",
                  "imgs/car.jpg",
                  "imgs/cheetah.jpg",
                  "imgs/banana.jpg",
                  "imgs/bus.jpg",
                  "imgs/parfum.jpg",
                  "imgs/alligator.jpg",
                  "imgs/arc.jpg"]).launch()