bmmalone

[core]
    analytics = false
    remote = remote_storage
[cache]
    dir = /srv/dvc-cache
    shared = group
    type = symlink
['remote "remote_storage"']
    url = gs://my-dvc-shared-bucket/dvc-cache-folder

bmmalone

{
    "editor.rulers": [
        88,
        120
    ],
    "editor.formatOnSave": true,
    "python.formatting.provider": "black",
    "python.formatting.blackArgs": [
        "--line-length",
        "120"

bmmalone

|-- LICENSE (if relevant)
|
|-- README.md			<- the top-level README for developers using this project
|
|-- CHANGELOG.md		<- a changelog for tracking changes over the history of a project.
|						   Example: https://github.com/bmmalone/pyllars/blob/dev/CHANGELOG.md
|
|-- setup.cfg			<- project-specific configuration for python.
|
|-- .gitignore			<- avoid uploading data, credentials, outputs, system files, etc.

bmmalone

#! /usr/bin/env bash

set -vx

gh_token=/path/to/my/gh-token.txt
account=my_account

# login to github
gh auth login --with-token < "$gh_token"

bmmalone

SHELL=/bin/bash
BASH_ENV=/path/to/my/.bashrc_conda

00 03 * * * cd /my/working/dir; conda activate my_env; /path/to/my/script.py >>/path/to/my.stdout 2>>/path/to/my.stderr; conda deactivate

bmmalone

# check running kernel version
uname -r

# check installed kernel headers
apt list --installed | grep "linux-headers"

# if the correct ones aren't installed, then install them
sudo apt install linux-headers-$(uname -r)

# probably reboot, but I'm not sure if it's necessary

bmmalone

#! /usr/bin/env bash

image_loc="/data/docker"

###
# https://docs.docker.com/engine/install/debian/
###

# set up the docker repository

bmmalone

# Configuration file for jupyter-notebook.

def add_home_dir(*fn):
    import os.path
    return os.path.join(os.path.expanduser('~'), *fn)


#------------------------------------------------------------------------------
# Configurable configuration
#------------------------------------------------------------------------------

bmmalone

given labeled_training_indices (e.g., maybe there are 20 labeled training instances)
given labeled_test_indices (there are always ~3000 of these due to the split created by Harutyunyan et al.)

train_fold, val_fold <- stratified split(labeled_training_indices, train=70%, "test"=30%) # "test" is really the validation set here

# for example, if we have 20 labeled training instances, then we have 14 instances for training and 6 for validation
# ... so we really don't have a lot when the number of labeled training instances is small

hp_grid = ParameterGrid({
	'penalty': ['l1', 'l2'],

bmmalone

#! /usr/bin/env bash

function mkipycondaenv {

    if [ "$#" -ne 1 ]; then
        echo "usage: mkipycondaenv <condaenv_name>"
        return 1
    fi

    conda create --name "$1" pip wheel jupyter ipykernel