2021-12-14 14:02:19 +08:00
|
|
|
import requests
|
2022-01-21 21:44:12 +08:00
|
|
|
import torch
|
2021-12-14 14:02:19 +08:00
|
|
|
from PIL import Image
|
|
|
|
|
from torchvision import transforms
|
|
|
|
|
|
2022-01-21 21:44:12 +08:00
|
|
|
import gradio as gr
|
|
|
|
|
|
|
|
|
|
model = torch.hub.load("pytorch/vision:v0.6.0", "resnet18", pretrained=True).eval()
|
2021-12-14 14:02:19 +08:00
|
|
|
|
|
|
|
|
# Download human-readable labels for ImageNet.
|
|
|
|
|
response = requests.get("https://git.io/JJkYN")
|
|
|
|
|
labels = response.text.split("\n")
|
|
|
|
|
|
2022-01-21 21:44:12 +08:00
|
|
|
|
2021-12-14 14:02:19 +08:00
|
|
|
def predict(inp):
|
2022-01-21 21:44:12 +08:00
|
|
|
inp = Image.fromarray(inp.astype("uint8"), "RGB")
|
|
|
|
|
inp = transforms.ToTensor()(inp).unsqueeze(0)
|
|
|
|
|
with torch.no_grad():
|
|
|
|
|
prediction = torch.nn.functional.softmax(model(inp)[0], dim=0)
|
|
|
|
|
return {labels[i]: float(prediction[i]) for i in range(1000)}
|
|
|
|
|
|
2021-12-14 14:02:19 +08:00
|
|
|
|
|
|
|
|
inputs = gr.inputs.Image()
|
|
|
|
|
outputs = gr.outputs.Label(num_top_classes=3)
|
|
|
|
|
gr.Interface(fn=predict, inputs=inputs, outputs=outputs).launch()
|
