This example is a 4-dimensional geometric brownian motion. The code
for the torchsde version is pulled directly from the
torchsde README
so that it would be a fair comparison against the author's own code.
The only change to that example is the addition of a dt choice so that
the simulation method and time step matches between the two different programs.
The SDE is solved 100 times. The summary of the results is as follows:
| using PyFenicsAD | |
| using Zygote | |
| using PyCall | |
| using Turing | |
| import LinearAlgebra: norm | |
| using Random, Distributions | |
| fenics = pyimport("fenics") | |
| fenics.set_log_level(fenics.LogLevel.ERROR) | |
| fa = pyimport("fenics_adjoint") |
| import numpy as np | |
| import math | |
| import torch | |
| from torchvision import datasets | |
| import cv2 # OpenCV, this is only used for visualization, see bottom of file | |
| def rotation_matrix(axis, theta): | |
| """ | |
| Generalized 3d rotation via Euler-Rodriguez formula, https://www.wikiwand.com/en/Euler%E2%80%93Rodrigues_formula |
| # I used the commands below not only on the command line, but also set up the aliased commands as global keyboard shortcuts. | |
| # This can be useful if you have a keyboard that lacks media keys, or if the media keys don't behave exactly as you want them to. | |
| # Everything below assumes PulseAudio on top of ALSA. | |
| alias volup="amixer set 'Master' 10%+" # Using amixer because pacmd/pactl can increase beyond 100% volume, which I don't want. | |
| alias voldn="amixer set 'Master' 10%-" | |
| # This assumes sink port #1 and the naming of these ports. Check `pacmd list-sinks` and make adjustments if necessary. | |
| alias outsp='pacmd "set-sink-port 1 analog-output;output-speaker' # output to speakers | |
| alias outhp='pacmd "set-sink-port 1 analog-output;output-headphones-1' # output to headphones |
on linux 4.19, I get
error creating new backup file '/var/lib/dpkg/status-old': Invalid cross-device link
when trying to install stuff on a ubuntu docker container. this seems to be an issue with metacopy https://www.spinics.net/lists/linux-unionfs/msg06109.html
issue can be worked around by running
| /* eslint-disable */ | |
| const common = require('../jest-config.common'); | |
| module.exports = Object.assign({}, common, { | |
| "coverageDirectory": "<rootDir>/coverage/browser", | |
| "testMatch": [ | |
| "**/test/browser/**/*.test.js" | |
| ], | |
| "setupFiles": [ | |
| "<rootDir>/src/test/browser/setup.js" | |
| ] |
Here's a simple implementation of bilinear interpolation on tensors using PyTorch.
I wrote this up since I ended up learning a lot about options for interpolation in both the numpy and PyTorch ecosystems. More generally than just interpolation, too, it's also a nice case study in how PyTorch magically can put very numpy-like code on the GPU (and by the way, do autodiff for you too).
For interpolation in PyTorch, this open issue calls for more interpolation features. There is now a nn.functional.grid_sample() feature but at least at first this didn't look like what I needed (but we'll come back to this later).
In particular I wanted to take an image, W x H x C, and sample it many times at different random locations. Note also that this is different than upsampling which exhaustively samples and also doesn't give us fle
Elm is a statically typed functional language that compiles to JavaScript. It's well-known for its developer experience: the compiler provides nice error messages, the package system enforces semantic versioning for all published packages and makes sure every exposed value or type has some documentation and type annotations.
| import Image | |
| import sys | |
| import glob | |
| import ImageOps | |
| # Trim all png images with white background in a folder | |
| # Usage "python PNGWhiteTrim.py ../someFolder" | |
| try: | |
| folderName = sys.argv[1] |