Yassine Alouini yassineAlouini

⚙️

PyTorch Exploration...

A data scientist that enjoys learning and experimenting with new tools. Focusing on computer vision applications for now.

yassineAlouini / denoise.py

Created December 17, 2023 16:30

A short script to denoise the audio of a noisy video.

	import soundfile as sf
	import numpy as np
	import librosa
	import noisereduce as nr

	# If you don't have the libraries above:
	# pip install noisereduce librosa numpy soundfile

	# Run ffmpeg first to extract the sound from the video
	# ffmpeg -i 'noisy_video.mp4' -vn -acodec copy 'noisy_signal.aac'

yassineAlouini / day15.py

Created February 2, 2022 09:31

Solution for day 15 AoC 2021

	def main():
	import networkx as nx
	import numpy as np
	data = np.loadtxt("/home/yassinealouini/Documents/code/advent_of_code/aoc/year_2021/data/15.txt", "U")
	size = int(len(data[0]))
	data = data.view('U1').astype(int).reshape(data.shape[0], size)
	# Need digraph since moving from previous to next position is not the same as the reverse.
	G = nx.DiGraph()

	def new_grid(data):

yassineAlouini / bfs_shortest_path.py

Created January 1, 2022 14:48

Find shortest path in an unweighted undirected graph using BFS

	# Simple bfs implementation to explore a graph.
	# An undirected graph as an adjancency representation.


	from collections import deque

	G = {0: (1, 2), 1: (0, 6), 6: (1, 3), 3: (2, 4, 6), 5: (2, ), 2: (5, 0, 3), 4: (3, )}
	source = 6
	end = 2

yassineAlouini / aoc_24_2021.py

Created December 24, 2021 15:54

Solution to AOC 2021 Day 24

	import math
	import numpy as np
	data = open("/home/yassinealouini/Documents/code/advent_of_code/aoc/year_2021/data/24.txt").read().rstrip()


	OPERATIONS = {"add": lambda x: sum(x), "mul": lambda x: math.prod(x),
	"div": lambda x: x[0] // x[1], "mod": lambda x: x[0] % x[1],
	"inp": lambda x: x[1],
	"eql": lambda x: int(x[0] == x[1])}

yassineAlouini / install_cuda.sh

Created February 27, 2021 09:42

In case your latest Ubuntu update breaks everything for CUDA. :p

	# From https://heads0rtai1s.github.io/2021/02/25/gpu-setup-r-python-ubuntu/


	# Fresh install
	sudo apt-get --purge remove "cublas" "cufft" "curand" "cusolver" "cusparse" "npp" "nvjpeg" "cuda" "nsight"
	sudo apt-get --purge remove "nvidia"
	sudo apt-get autoremove

	# Drivers
	wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64/cuda-ubuntu2004.pin

yassineAlouini / model.py

Created January 19, 2021 14:44

Bad QA code

	from sklearn import svm
	from sklearn.datasets import load_iris
	from joblib import dump
	from joblib import load
	import sys
	import os
	import pathlib
	import pandas as pd

	TRAIN = False

yassineAlouini / total_variations.py

Created February 14, 2020 09:30

Compute the variations of the total column

	def get_input_df():

	df = pd.DataFrame({"date": pd.date_range("2017-1-1", "2020-1-1", freq="1D")})
	df["open"] = np.random.choice([1, 0], len(df))
	df["closed"] = np.random.choice([1, 0], len(df))
	df["location"] = np.random.choice(["a", "b", "c"], len(df))
	return df

	def compute_total_variations_df(df):
	df = df.copy()

yassineAlouini / remote_jupyter.sh

Created November 4, 2019 09:33

Remote Jupyter access to my desktop aliases

	alias port_forward='nohup ssh -N -f -L localhost:8889:localhost:8889 deepmeaning'
	alias remote_notebook_start='nohup ssh -f deepmeaning "jupyter notebook --no-browser --port=8889"; port_forward'
	alias remote_notebook_stop='ssh deepmeaning "pkill -u yassinealouini jupyter"'

yassineAlouini / df_csv_bz2_to_s3.py

Created February 28, 2019 15:58

Save DataFrame to S3 as a CSV with bz2 compression

	import s3fs
	import pandas as pd
	import bz2file
	s3 = s3fs.S3FileSystem()

	df = pd.DataFrame() # Replace with your DataFrame
	dest_s3_key = "" # Where to save in S3

	with s3.open(dest_s3_key, 'wb') as dest_file:
	with bz2file.open(dest_file, 'wb') as compressed_file:

yassineAlouini / upload_image.html

Created February 24, 2019 13:26

A simple image uploading HTML form

	<!doctype html>
	<title>Upload</title>
	<h1>Upload an image to recognize</h1>
	<form method=post enctype=multipart/form-data action="{{url_for('upload_img')}}">
	<p><input type=file name=file> <input type=submit value=Upload> </p>
	</form>