SHAP_batch_gpu1.py

import torch, torchvision
from torch import nn
from torchvision import transforms, models, datasets
import shap
import json
import numpy as np
from matplotlib import pyplot as plt
import os
import time
from srblib import abs_path
from PIL import ImageFilter, Image
import torch.utils.data
import torch.utils.data.distributed
import torchvision.transforms as transforms
import torchvision.models as models
from tqdm import tqdm
import shutil
import skimage
from skimage.transform import resize

val_dir = './val'

imagenet_val_xml_path = './val_bb'
imagenet_val_path = './val/'
base_img_dir = abs_path(imagenet_val_path)
input_dir_path = 'images_list.txt'
text_file = abs_path(input_dir_path)
imagenet_class_mappings = './imagenet_class_mappings'

img_name_list = []
with open(text_file, 'r') as f:
    for line in f:
        img_name_list.append(line.split('\n')[0])


def imagenet_label_mappings():
    fileName = os.path.join(imagenet_class_mappings, 'imagenet_label_mapping')
    with open(fileName, 'r') as f:
        image_label_mapping = {int(x.split(":")[0]): x.split(":")[1].strip()
                               for x in f.readlines() if len(x.strip()) > 0}
        return image_label_mapping


class DataProcessing:
    def __init__(self, data_path, transform, img_idxs=[0, 1], if_noise=0, noise_var=0.0):
        self.data_path = data_path
        self.transform = transform
        self.if_noise = if_noise
        self.noise_mean = 0
        self.noise_var = noise_var

        img_list = img_name_list[img_idxs[0]:img_idxs[1]]
        self.img_filenames = [os.path.join(data_path, f'{i}.JPEG') for i in img_list]


    def __getitem__(self, index):
        img = Image.open(os.path.join(self.data_path, self.img_filenames[index])).convert('RGB')
        target = self.get_image_class(os.path.join(self.data_path, self.img_filenames[index]))

        if self.if_noise == 1:
            img = skimage.util.random_noise(np.asarray(img), mode='gaussian',
                                            mean=self.noise_mean, var=self.noise_var,
                                            )  # numpy, dtype=float64,range (0, 1)
            img = Image.fromarray(np.uint8(img * 255))
            # print(img2.max(), img2.min())

        img = self.transform(img)
        return img, target, os.path.join(self.data_path, self.img_filenames[index])
        # return img, target

    def __len__(self):
        return len(self.img_filenames)

    def get_image_class(self, filepath):
        # ImageNet 2012 validation set images?
        with open(os.path.join(imagenet_class_mappings, "ground_truth_val2012")) as f:
            ground_truth_val2012 = {x.split()[0]: int(x.split()[1])
                                    for x in f.readlines() if len(x.strip()) > 0}

        def get_class(f):
            ret = ground_truth_val2012.get(f, None)
            return ret

        image_class = get_class(filepath.split('/')[-1])
        return image_class


transform_val = transforms.Compose([
    transforms.Resize((256, 256)),
    transforms.CenterCrop(224),
    transforms.ToTensor(),
    transforms.Normalize(mean=[0.485, 0.456, 0.406],
                         std=[0.229, 0.224, 0.225]),
])


# Plots image from tensor
def tensor_imshow(inp, title=None, **kwargs):
    """Imshow for Tensor."""
    inp = inp.numpy().transpose((1, 2, 0))
    # Mean and std for ImageNet
    mean = np.array([0.485, 0.456, 0.406])
    std = np.array([0.229, 0.224, 0.225])
    inp = std * inp + mean
    inp = np.clip(inp, 0, 1)
    plt.imshow(inp, **kwargs)
    if title is not None:
        plt.title(title)
    plt.axis('off')
    plt.show()


init_time = time.time()
val_dataset = DataProcessing(base_img_dir, transform_val, img_idxs=[100, 200], if_noise=0, noise_var=0.0)
val_loader = torch.utils.data.DataLoader(val_dataset, batch_size=50, shuffle=False, num_workers=24, pin_memory=True)

# especificar cual gpu 0 o 1
torch.cuda.set_device(1)
# load the model
modelo = 'alexnet'
# load the model
if modelo == 'googlenet':
    model = models.googlenet(pretrained=True)
elif modelo == 'vgg16':
    model = models.vgg16(pretrained=True)
elif modelo == 'resnet50':
    model = models.resnet50(pretrained=True)
elif modelo == 'alexnet':
    model = models.alexnet(pretrained=True)
model.cuda()
model.eval()

im_label_map = imagenet_label_mappings()

iterator = tqdm(enumerate(val_loader), total=len(val_loader), desc='batch')

save_path = './{}_SHAP'.format(modelo)

for i, (images, target, path) in iterator:
    images = images.cuda()
    if modelo == 'googlenet':
        e = shap.GradientExplainer((model, model.conv2), images)
    elif modelo == 'resnet50':
        e = shap.GradientExplainer((model, model.layer1[0].conv1), images)
    elif modelo == 'vgg16':
        e = shap.GradientExplainer((model, model.features[7]), images)
    elif modelo == 'alexnet':
        e = shap.GradientExplainer((model, model.features[3]), images)

    shap_values, indexes = e.shap_values(images, ranked_outputs=1, nsamples=200)

    heatmap = np.clip(shap_values[0].sum(1), 0, 1)

    for j, file_name in enumerate(path):
        mask_file = ('{}_mask.npy'.format(file_name.split('/')[-1].split('.JPEG')[0]))
        np.save(os.path.abspath(os.path.join(save_path, mask_file)), resize(heatmap[j], (224, 224)))
        # plt.imshow(heatmap[j])
        # plt.show()

print('Time taken: {:.3f}'.format(time.time() - init_time))