zommiommy

import time
import numpy as np
from tqdm.auto import trange
import matplotlib.pyplot as plt
from optimizers import SGD, Momentum, Adam, Nadam

# Generate some data
positives = np.random.normal(loc=-0.3, size=1000)
negatives = np.random.normal(loc=0.0, size=1000)

zommiommy

#![feature(portable_simd)]
#![feature(test)]

extern crate test;
use test::{black_box, Bencher};
use std::convert::TryInto;
use std::simd::*;
use std::arch::x86_64::*;

type Reg = f32x8;

zommiommy

[package]
name = "bencher"
version = "0.1.0"
edition = "2021"

# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html

[dependencies]
libc = "0.2"

zommiommy

test bench_naive_split                       ... bench:   2,265,338 ns/iter (+/- 136,512)
test bench_naive_split_no_escaping           ... bench:   2,287,644 ns/iter (+/- 309,993)
test bench_naive_split_no_escaping_hardcoded ... bench:   2,650,552 ns/iter (+/- 27,290)
test bench_rust_default_split                ... bench:   4,118,575 ns/iter (+/- 153,705)
test bench_simd_split_no_escaping            ... bench:   1,707,197 ns/iter (+/- 83,440)
test bench_simd_split_no_escaping_hardcoded  ... bench:   1,688,714 ns/iter (+/- 11,317)

zommiommy

build:
  nasm -f elf64 -o measure.o ./measure.asm
  ld measure.o -o measure.out -lc --dynamic-linker /lib/ld-2.33.so

zommiommy

M0lecon RE Writeup

TLDR

• Reverse the code
• Re-implment it in Rust
• Bruteforce
• ???
• Profit

Automatic Rejection Machine

zommiommy

from tensorflow.keras.layers import Input, Dense, Layer, Lambda
from tensorflow.keras.models import Model
import tensorflow.keras.backend as K

def setup_embrace_layer(modalities=2, num_samples=256):
    def embrace(dockings):
        batch_size = K.shape(dockings[0])[0]
        # Compute the probabilities of extraction
        selection_probabilities = tf.ones([batch_size, len(dockings)], dtype=tf.dtypes.float32)
        probabilty_sum = tf.reduce_sum(selection_probabilities, axis=-1, keepdims=True)  # [batch_size, 1]

zommiommy

from pwn import *

libc = ELF('/lib/x86_64-linux-gnu/libc-2.31.so')
host = '13.231.226.137'
port = 9573

p = remote(host, port)

def op():

zommiommy

import tensorflow as tf
import tensorflow.keras.backend as K
from tensorflow.keras.metrics import Metric, AUC
from tensorflow.keras.backend import epsilon

class ConfusionMatrixMetric(Metric):
    def __init__(self, name, **kwargs):
        super(ConfusionMatrixMetric, self).__init__(name=name, **kwargs)
        self.tp = self.add_weight(name='tp', initializer='zeros')
        self.fp = self.add_weight(name='fp', initializer='zeros')

zommiommy

How to setup tensorflow gor gpu on Ubuntu

To check if tensorflow can detect any GPU just run

python -c "import tensorflow as tf;print(tf.test.is_gpu_available(True))"

If the GPUs are available it should print True.

Setup anaconda

The easiest way to setup cuda and tensorflow-gpu is to install everything using anaconda. Anaconda do not requires root permissions and create a self-contained folder in $HOME/anaconda3.