Chao YIN (尹超) PointCloudYC

Steps to manage multiple CUDA environments

Latest Update: May 19th, 2024

This gist contains all the steps required to:

Install multiple CUDA versions (e.g., CUDA 11.8 and CUDA 12.1
Manage multiple CUDA environments on Ubuntu using the utility called environment modules.
Use this approach to avoid CUDA environment conflicts.

Environment Modules is a package that provides for the dynamic modification of a user's environment via modulefiles. You can find more on it at https://modules.readthedocs.io/en/latest/

Goal

Install working tensorflow or pytorch via standard conda environment workflow.

Basic Setup : Install `pytorch` in a fresh conda environment

The recommended conda-based install process works smoothly:

$ # Create a fresh environment

Installation

I remember setting up cuda 9.0 was quite challenging for me and I expected 10.0 to be no different However, to my surprise the installation was rather easy

Download

Download CUDA 10.0, I chose the following options:

Option	Value
Operating System	Linux

Download CMake from: https://cmake.org/download/

wget https://cmake.org/files/v3.12/cmake-3.12.3.tar.gz

Compile from source and install

tar zxvf cmake-3.*

Ways to remap caps lock to control on Windows 10

These methods in this gist worked for me on my U.S.-based keyboard layouts. I am unsure about other layouts. If you have problems, revert your changes; delete the registry key you created (and reboot).

Update: you should probably scroll down to approach 4 where I suggest using Microsoft PowerToys Keyboard Manager.

Approach 1. Manually through regedit

Navigate to and create a new binary value in HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Keyboard Layout named Scancode Map.

	def soft_dice_loss(y_true, y_pred, epsilon=1e-6):
	'''
	Soft dice loss calculation for arbitrary batch size, number of classes, and number of spatial dimensions.
	Assumes the `channels_last` format.

	# Arguments
	y_true: b x X x Y( x Z...) x c One hot encoding of ground truth
	y_pred: b x X x Y( x Z...) x c Network output, must sum to 1 over c channel (such as after softmax)
	epsilon: Used for numerical stability to avoid divide by zero errors

	#!/bin/bash
	#
	# script to extract ImageNet dataset
	# ILSVRC2012_img_train.tar (about 138 GB)
	# ILSVRC2012_img_val.tar (about 6.3 GB)
	# make sure ILSVRC2012_img_train.tar & ILSVRC2012_img_val.tar in your current directory
	#
	# https://github.com/facebook/fb.resnet.torch/blob/master/INSTALL.md
	#
	# train/

	#!/bin/bash

	# Script for installing tmux on systems where you don't have root access.
	# tmux will be installed in $HOME/local/bin.
	# It's assumed that wget and a C/C++ compiler are installed.

	# exit on error
	set -e

	TMUX_VERSION=2.3

	"""
	A weighted version of categorical_crossentropy for keras (2.0.6). This lets you apply a weight to unbalanced classes.
	@url: https://gist.github.com/wassname/ce364fddfc8a025bfab4348cf5de852d
	@author: wassname
	"""
	from keras import backend as K
	def weighted_categorical_crossentropy(weights):
	"""
	A weighted version of keras.objectives.categorical_crossentropy

Chao YIN (尹 超) PointCloudYC