Marcel Santana marcelsan

💭

I may be slow to respond.

Deep Learning applied to Computer Graphics, Computational Photography and Computer Vision.

marcelsan / midas_loss.py

Created October 9, 2020 14:40 — forked from dvdhfnr/midas_loss.py

Loss function of MiDaS

	import torch
	import torch.nn as nn


	def compute_scale_and_shift(prediction, target, mask):
	# system matrix: A = [[a_00, a_01], [a_10, a_11]]
	a_00 = torch.sum(mask * prediction * prediction, (1, 2))
	a_01 = torch.sum(mask * prediction, (1, 2))
	a_11 = torch.sum(mask, (1, 2))

marcelsan / vgg.py

Created February 5, 2019 15:23

This script will demonstrate how to use a pretrained model, in PyTorch, to make predictions.

	"""
	This script will demonstrate how to use a pretrained model, in PyTorch,
	to make predictions. Specifically, we will be using VGG16 with a cat
	image.

	References used to make this script:
	PyTorch pretrained models doc:
	http://pytorch.org/docs/master/torchvision/models.html
	PyTorch image transforms example:
	http://pytorch.org/tutorials/beginner/data_loading_tutorial.html#transforms

marcelsan / kde.py

Created December 9, 2018 00:09

	import numpy as np

	class KernelDensity:
	def __init__(self, kernel='gaussian', bandwidth=0.2):
	self.kernel_ = kernel
	self.bandwidth_ = bandwidth

	def fit(self, X):
	self.X_ = X
	self.n_ = X.shape[0]

marcelsan / disp_multiple_images.py

Created March 15, 2018 19:35 — forked from soply/disp_multiple_images.py

Plot multiple images with matplotlib in a single figure. Titles can be given optionally as second argument.

	import matplotlib.pyplot as plt
	import numpy as np

	def show_images(images, cols = 1, titles = None):
	"""Display a list of images in a single figure with matplotlib.

	Parameters
	---------
	images: List of np.arrays compatible with plt.imshow.

marcelsan / pca_intrinsic_dimension.py

Last active January 4, 2018 02:18

	from sklearn.decomposition import PCA
	import matplotlib.pyplot as plt
	from sklearn import datasets

	iris = datasets.load_iris()
	X = iris.data

	model = PCA()
	model.fit(X)

marcelsan / jupyter.sh

Created December 26, 2017 17:31 — forked from Colelyman/jupyter.sh

Get Jupyter Notebook Running in SSH Tunnel

	user@server$ jupyter notebook --no-browser --port=8889
	# run client side
	user@client$ ssh -N -f -L localhost:8000:localhost:8889 remote_user@remote_host
	firefox http://localhost:8000
	# source: https://coderwall.com/p/ohk6cg/remote-access-to-ipython-notebooks-via-ssh

marcelsan / Range.hpp

Created November 8, 2017 17:31 — forked from tunabrain/Range.hpp

Python ranges in C++

	#ifndef RANGE_HPP_
	#define RANGE_HPP_

	template<typename T> class RangeIterator;

	template<typename T>
	class Range
	{
	T _start, _end, _step;