jeremy-rutman’s gists

jeremy-rutman / LI_page_source.html

Created March 26, 2021 14:22

LI_page_source

This file has been truncated, but you can view the full file.

	<html class="theme theme--mercado artdeco" lang="en"><head>
	<script async="" src="https://platform.linkedin.com/js/analytics.js"></script><script type="application/javascript">!function(i,n){void 0!==i.addEventListener&&void 0!==i.hidden&&(n.liVisibilityChangeListener=function(){i.hidden&&(n.liHasWindowHidden=!0)},i.addEventListener("visibilitychange",n.liVisibilityChangeListener))}(document,window);</script>

	<title>(7) Assaf Govari, Ph.D. \| LinkedIn</title>

	<meta charset="utf-8">
	<meta http-equiv="X-UA-Compatible" content="IE=edge">
	<meta name="asset-url" class="mercado-icons-sprite" id="artdeco-icons/static/images/sprite-asset" content="https://static-exp1.licdn.com/sc/h/brlx1tn7p1njjhu3u6kjxqodw">
	<meta name="description" content="">
	<meta name="google" content="notranslate">

jeremy-rutman / venn_diagram_with_three_circles

Last active February 14, 2021 10:34

venn diagram with three circles showing overlap order

	from matplotlib_venn import venn2, venn3
	import matplotlib.pyplot as plt

	# venn2 subsets are a b ab (where a is n_unique_to_a, ab is intersection of a,b)
	# called by venn2(subsets = (a_name,b_name,intersection_name), set_labels = (pair[0], pair[1],'both'), alpha = 0.5)

	# venn3 subsets are a b ab c ca cb abc
	# called by venn3(subsets = (a,b,ab,c,ca,cb,n),
	# set_labels = (a_name,b_name,c_name), alpha = 0.5)

jeremy-rutman / pdf_color_to_gray.sh

Last active November 3, 2020 19:46

convert color pdf to grayscale

	jeremy@jeremy-Blade:$ gs \
	-sOutputFile=output.pdf \
	-sDEVICE=pdfwrite \
	-sColorConversionStrategy=Gray \
	-dProcessColorModel=/DeviceGray \
	-dCompatibilityLevel=1.4 \
	-dNOPAUSE \
	-dBATCH \
	input.pdf

jeremy-rutman / torchtext_info.txt

Created June 25, 2020 09:14

	Some notes on torchtext

	you can read a csv and generate vocab like this

	tokenize = lambda x: str(x).split() # see if this fixes float vs. string error
	TEXT = data.Field(sequential=True, tokenize=tokenize, lower=True, include_lengths=True, batch_first=True,
	fix_length=200)

	LABEL = data.LabelField() # LABEL = data.LabelField(tensor_type=torch.FloatTensor)
	fields = {'textcol': ('text', TEXT), 'real': ('label',LABEL)}

jeremy-rutman / pandas_tricks.py

Last active June 24, 2020 10:21

	# selecting all cols except one
	df = pd.DataFrame({'a':[1,2,3,4],'b':[1,2,3,4]})
	df2 = df.loc[:,df.columns!='b']
	print(df)
	print(df2)

	# split df into train, val, test with val from 0.9 to 0.95 and test from 0.95 to 1.0 of randomized data
	train, validate, test = np.split(df.sample(frac=1), [int(.9len(df)), int(.95len(df))])

	# FILTERING

jeremy-rutman / test_batchwise_loss.py

Last active March 9, 2020 11:17

	# attempt at 'batch-known' learning , where positive/negative is known per batch and not per example
	# in this attempt, examples are positive if there are five or more ones in the input vector which is itself six random binary digits.
	# not only is this underlying truth (y=(sum(x)>=5)) hidden from the loss function , the loss func will only know per batch if there was at
	# least one true example in the batch (a 'positive batch'), or none (a 'negative batch')


	import pandas as pd
	import matplotlib.pyplot as plt
	from mpl_toolkits.mplot3d import Axes3D
	import torch

jeremy-rutman / nvidia_nightmare

Last active May 26, 2025 20:36

	tl;dr
	use the package manager install (.deb or .rpm files, not .run file)

	remove stuff using

	sudo apt-get --purge remove 'cuda*'
	sudo apt-get --purge -y remove 'nvidia*'
	sudo apt-get --purge -y remove 'libnvidia*'

	using

jeremy-rutman / geyser_gazer.py

Last active January 8, 2020 15:54

geyser data analysis

	# code for https://unclejerry9466728.wordpress.com/2019/12/28/wiser-geyser/

	import pandas as pd
	import matplotlib.pyplot as plt
	from mpl_toolkits.mplot3d import Axes3D
	import torch
	import numpy as np
	import pandas as pd

jeremy-rutman / image_warp.py

Last active December 19, 2019 12:58

warp image arbitrarily (nonlinear)

	import numpy as np
	import cv2

	def remap_test(img_arr):
	H,W = img_arr.shape[0:2]
	map = np.mgrid[0:H,0:W] # the map doesnt have to be same dims. as image size; it just declares what pixels in input image map to given output pixel
	map_x = map[1].astype(np.float32)
	map_y = map[0].astype(np.float32)
	#linear warp
	# for i in range(H):

jeremy-rutman / change_LUT.py

Created December 10, 2019 08:26

from 2d values to RGB

	import cv2
	import numpy as np
	def get_rgb(two_d_pixels):
	theta = np.arctan(two_d_pixels[1,:] / two_d_pixels[0,:])
	r = np.sqrt(two_d_pixels[1,:]2 + two_d_pixels[0,:]2)
	hue = (theta+3.14/2) * 128./(3.14) # 127 max for hue
	sat = (r/np.sqrt(2)) * 255 #255 max for sat
	H,W = two_d_pixels.shape[1:3]
	retval = np.zeros([H,W,3],dtype = np.uint8)
	retval[:,:,0] = hue

jeremy rutman jeremy-rutman