antoinebrl

import torch
from torch import nn
from torch.utils.data import DataLoader
from torchvision import datasets
from torchvision.transforms import ToTensor, Lambda, Compose

transform = Compose([ToTensor(), Lambda(lambda x: torch.flatten(x))])
train_dataset = datasets.MNIST(root='data', train=True, transform=transform, download=True)
test_dataset = datasets.MNIST(root='data', train=False, transform=transform)
train_loader = DataLoader(train_dataset, batch_size=250, shuffle=True, drop_last=True)

antoinebrl

import timm
import torchvision
import torch
import urllib
from PIL import Image
from torch import nn
import onnx
import onnx.helper as oh
from onnx.onnx_pb import ValueInfoProto
import numpy as np

antoinebrl

# Images have color pixels encoded as uint8. In computer
# vision and deep learning the images is represented with
# floatingpoints with value in the range [0,1]. This is
# obtained by dividing by 255, the highest color intensity.
# However, reduced precision training an inference become
# the new standard.
#
# What is the magnitude of the rouding error when normalizing
# an image by 255 rather than 256 (256=2^8)?

antoinebrl

import numpy as np
from scipy.stats import mode


def euclidean_distance(x1, x2):
    # x1 array of shape (N, D)
    # x2 array of shape (N', D)
    # output pairwise distances, array of shape (N, N')
    return np.linalg.norm(x1[:, np.newaxis] - x2, axis=1)

antoinebrl

find . ! -user root -exec chown root:root {} \;
find . -type f ! -perm 644 -exec chmod 644 {} \;
find . -type d ! -perm 755 -exec chmod 755 {} \;

antoinebrl

import onnx

model = onnx.load("model.onnx")
graph = model.graph

# Identify relevant node
idx = -1
for i, n in enumerate(model.graph.node):
    if n.name == "GlobalAveragePool":
        idx = i

antoinebrl

Preparation of ImageNet (ILSVRC2012)

The dataset can be found on the official website if you are affiliated with a research organization. It is also available on Academic torrents.

This script extracts all the images and group them so that folders contain images that belong to the same class.

1. Download the ILSVRC2012_img_train.tar and ILSVRC2012_img_val.tar
2. Download the script wget https://gist.githubusercontent.com/antoinebrl/7d00d5cb6c95ef194c737392ef7e476a/raw/dc53ad5fcb69dcde2b3e0b9d6f8f99d000ead696/prepare.sh
3. Run it ./prepare.sh

antoinebrl

Sources

• Google XSS Challenges at https://xss-game.appspot.com
• Help : https://www.google.com/about/appsecurity/learning/xss/index.html

Level 1: Hello, world of XSS

The output will be directly displayed without any escaping. A simple query with <script>alert("XSS")</script> will do the job.