cnn_utils.py

import torch
import torch.cuda


def evaluate_single(input_individual, input_model, data_loader, device):
    correct = 0
    total = 0
    input_model = input_model.eval()
    input_model.set_individual(input_individual)
    with torch.no_grad():
        for data in data_loader:
            images, labels = data
            images = images.to(device)
            labels = labels.to(device)
            outputs = input_model(images)
            _, predicted = torch.max(outputs[0].data, 1)
            total += labels.size(0)
            correct += (predicted == labels).sum().item()
    return 100 * correct / total


def evaluate_individual_list(input_individual_list, ga, input_model, data_loader, device):
    correct = 0
    total = 0
    input_model = input_model.eval()
    i = 0
    with torch.no_grad():
        while len(input_individual_list) > i:
            for data in data_loader:
                if len(input_individual_list) <= i:
                    pass
                else:
                    ind = input_individual_list[i]
                    input_model.set_individual(ind)
                    images, labels = data
                    images = images.to(device)
                    labels = labels.to(device)
                    outputs = input_model(images)
                    _, predicted = torch.max(outputs[0].data, 1)
                    total += labels.size(0)
                    correct += (predicted == labels).sum().item()
                    acc = 100 * correct / total
                    ga.update_current_individual_fitness(ind, acc)
                    i += 1