app.py

import torch
import gradio as gr
from PIL import Image
from src.model import get_model, apply_weights, copy_weight
from src.transform import crop, pad, gpu_crop
from torchvision.transforms import Normalize, ToTensor
from pathlib import Path

vocab = [
    "Actinic Keratosis",
    "Basal Cell Carcinoma",
    "Benign Keratosis",
    "Dermatofibroma",
    "Melanoma",
    "Melanocytic Nevus",
    "Vascular Lesion",
]


model = get_model()
state = torch.load("exported_model.pth", map_location="cpu")
apply_weights(model, state, copy_weight)

to_tensor = ToTensor()
norm = Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])


def classify_image(inp):
    inp = Image.fromarray(inp)
    transformed_input = pad(crop(inp, (460, 460)), (460, 460))
    transformed_input = to_tensor(transformed_input).unsqueeze(0)
    transformed_input = gpu_crop(transformed_input, (224, 224))
    transformed_input = norm(transformed_input)
    model.eval()
    with torch.no_grad():
        pred = model(transformed_input)
    prob = torch.softmax(pred[0], dim=0)
    confidences = {vocab[i]: float(prob[i]) for i in range(7)}
    return confidences


iface = gr.Interface(
    fn=classify_image,
    inputs="image",
    outputs=gr.Label(),
    examples=[
        ["ISIC_0024634_00.jpg"],
        ["ISIC_0032932_00.jpg"],
    ],
    title="Skin Lesion Recognition using fast.ai",
    description="Adapted from https://domingomery.ing.puc.cl/",
    article="<p style='text-align: center'><a href='https://evertoncolombo.github.io/blog/posts/skin-lesion/Skin%20Lesion%20Recognition%20using%20fastai.html'>More info | <a href='https://www.dropbox.com/s/nzrvuoos7sgl5dh/exp4val.zip' >Dataset</a> <center><img src='https://visitor-badge.glitch.me/badge?page_id=e_colombo_skin_lesion' alt='visitor badge'></center></p>",
    allow_flagging="never",
).launch()