utils_eval.py

import torch
import torch.nn.functional as F
from tqdm import tqdm
from utils_loss_metric import dice_coeff


@torch.inference_mode()
def evaluate(net, dataloader, device, amp):
    net.eval()
    num_val_batches = len(dataloader)
    dice_score = 0

    for batch in tqdm(dataloader, total=num_val_batches, desc='Validation round', unit='batch', leave=False):
        image, mask_true = batch

        # move images and labels to correct device and type
        image = image.to(device=device, dtype=torch.float32)
        mask_true = mask_true.to(device=device, dtype=torch.long)

        # predict the mask
        mask_pred = net(image)
        mask_pred = (F.sigmoid(mask_pred) > 0.5).float()
        dice_score += dice_coeff(mask_pred, mask_true, reduce_batch_first=False)

    net.train()
    return dice_score / max(num_val_batches, 1)