ML in a Box

This repo houses a script used to turn any Paperspace Ubuntu 22.04 based Linux instance into a fully functional machine learning environment for interactive development. Users are free to modify and run against their own Paperspace instance, be it Linux headless or Linux desktop. The only requirement is a working Nvidia GPU driver pre-installed. A pre-built template exists in the Paperspace eco-system as a Public Template based on a base Ubuntu 22.04 image.

Who is this for?

We assume a generic advanced data science user who probably wants GPU access, but not any particular specialized subfield of data science such as computer vision or natural language processing. Such users can build upon this base to create their own stack, or we can create other VMs for subfields, similar to what can be done with Gradient containers.

Software included

Category	Software	Version	Install Method	Why / Notes
System	Nvidia Driver	535.129.03	pre-installed	Enable Nvidia GPUs. Latest version as of VM creation date
	CUDA	12.1.1	Apt	Nvidia A100 GPUs require CUDA 11+ to work, so 10.x is not suitable
	CUDA toolkit	12.1.1	Apt	Needed to work with Nvidia driver
	cuDNN	8.9.3.*-1+cuda12.1	Apt	Additional libarary to enhance CUDA functionality
	Python	3.11.6	Apt	Most widely used programming language for data science
	pip3	23.3.1	Apt	Enable easy installation of 1000s of other data science, etc., packages.
ML Frameworks	PyTorch	2.1.1	pip3	Most widely used deep learning framework
	Torchvision	0.16.1	pip3	Vision libary for PyTorch
	Torchaudio	2.1.1	pip3	Audio libary for PyTorch
	TensorFlow	2.15.0	pip3	Popular deep learning framework
Hugging Face	Transformers	4.35.2	pip3	Popular deep learning library for NLP brought to you by Hugging Face
	Datasets	2.14.5	pip3	A supporting Hugging Face library for datasets and data handling
	Peft	0.6.2	pip3	A Hugging Face Parameter-Efficient Fine-Tuning (PEFT) enables efficient adaptation of pre-trained language models to various downstream applications without fine-tuning all the model's parameters
	Tokenizers	0.13.3	pip3	A Hugging Face library supporting implementations of tokenizers
	Accelerate	0.24.1	pip3	A Hugging Face library used to support model training by abstracting boilerplate code
	Diffusers	0.21.4	pip3	A Hugging Face library used for implementation of diffusion models
	Safetensors	0.4.0	pip3	A Hugging Face library to store tensors safely
Supporting Libraries	JupyterLab	3.6.5	pip3	De facto standard for data science using Jupyter notebooks
	BitsandBytes	0.41.2	pip3	A lightweight wrapper around CUDA custom functions, in particular 8-bit optimizers
	Cloudpickle	2.2.1	pip3	Makes it possible to serialize Python constructs not supported by the default pickle module
	Scikit-image	0.21.0	pip3	Collection of algorithms for image processing
	Scikit-learn	1.3.0	pip3	Widely used ML library for data science, generally for smaller data or models
	Matplotlib	3.7.3	pip3	Widely used plotting library in Python for data science, e.g., scikit-learn plotting requires it
	IPywidgets	8.1.1	pip3	Interactive HTML widgets for Jupyter notebooks and the IPython kernel
	Cython	3.0.2	pip3	Enables writing C extensions for Python
	tqdm	4.66.1	pip3	Fast, extensible progress meter
	gdown	4.7.1	pip3	Google drive direct download of big files
	XGBoost	1.7.6	pip3	An optimized distributed gradient boosting library
	Pillow	9.5.0	pip3	Python imaging library
	seaborn	0.12.2	pip3	Python visualization library based on matplotlib
	SQLAlchemy	2.0.21	pip3	Python SQL toolkit and Object Relational Mapper that gives application developers the full power and flexibility of SQL
	spaCy	3.6.1	pip3	library for advanced Natural Language Processing in Python and Cython
	nltk	3.8.1	pip3	Natural Language Toolkit (NLTK) is a Python package for natural language processing
	boto3	1.28.51	pip3	Amazon Web Services (AWS) Software Development Kit (SDK) for Python
	tabulate	0.9.0	pip3	Pretty-print tabular data in Python
	future	0.18.3	pip3	The missing compatibility layer between Python 2 and Python 3
	jsonify	0.5	pip3	Provides the ability to take a .csv file as input and outputs a file with the same data in .json format
	opencv-python	4.8.0.76	pip3	Includes several hundreds of computer vision algorithms
	pyyaml	5.4.1	pip3	YAML parser and emitter for Python
	Sentence Transformers	2.2.2	pip3	A ML framework for sentence, paragraph and image embeddings
	wandb	0.15.10	pip3	CLI and library to interact with the Weights & Biases API (model tracking)
	Deepspeed	0.10.3	pip3	A DL optimization library for PyTorch designed to train large distrubuted models with better parallelism
	CuPyCUDA12x	12.2.0	pip3	A NumPy/SciPy-compatible array library for GPU-accelerated computing with Python
	timm	0.9.7	pip3	Deep-learning library that hosts a collection of SOTA computer vision models and tools
	OmegaConf	2.3.0	pip3	A hierarchical configuration system, with support for merging configurations from multiple sources
	SciPy	1.11.2	pip3	Fundamental algorithms for scientific computing in Python
	gradient	2.0.6	pip3	CLI and Python SDK for Paperspace Core and Gradient

Licenses

Software	License	Source
Nvidia Driver	Nvidia EULA	https://www.nvidia.com/en-us/drivers/nvidia-license/
Python	PSF	https://en.wikipedia.org/wiki/Python_(programming_language)
Pip3	MIT	https://github.com/pypa/pip/blob/main/LICENSE.txt
CUDA	NVidia EULA	https://docs.nvidia.com/cuda/eula/index.html
cuDNN	NVidia EULA	https://docs.nvidia.com/deeplearning/cudnn/sla/index.html
PyTorch	New BSD	https://github.com/pytorch/pytorch/blob/master/LICENSE
TensorFlow	Apache 2.0	https://github.com/tensorflow/tensorflow/blob/master/LICENSE
Transformers	Apache 2.0	https://github.com/huggingface/transformers/blob/main/LICENSE
Datasets	Apache 2.0	https://github.com/huggingface/datasets/blob/main/LICENSE
Peft	Apache 2.0	https://github.com/huggingface/peft/blob/main/LICENSE
Tokenizers	Apache 2.0	https://github.com/huggingface/tokenizers/blob/main/LICENSE
Accelerate	Apache 2.0	https://github.com/huggingface/accelerate/blob/main/LICENSE
Diffusers	Apache 2.0	https://github.com/huggingface/diffusers/blob/main/LICENSE
Safetensors	Apache 2.0	https://github.com/huggingface/safetensors/blob/main/LICENSE
JupyterLab	New BSD	https://github.com/jupyterlab/jupyterlab/blob/master/LICENSE
BitsandBytes	MIT	https://github.com/TimDettmers/bitsandbytes/blob/main/LICENSE
Cloudpickle	New BSD	https://github.com/cloudpipe/cloudpickle/blob/master/LICENSE
Scikit-image	New BSD	https://github.com/scikit-image/scikit-image/blob/main/LICENSE.txt
Scikit-learn	New BSD	https://github.com/scikit-learn/scikit-learn/blob/main/COPYING
Matplotlib	PSF-based	https://matplotlib.org/stable/users/license.html
IPywidgets	New BSD	https://github.com/jupyter-widgets/ipywidgets/blob/master/LICENSE
Cython	Apache 2.0	https://github.com/cython/cython/blob/master/LICENSE.txt
tqdm	MIT	https://github.com/tqdm/tqdm/blob/master/LICENCE
gdown	MIT	https://github.com/wkentaro/gdown/blob/main/LICENSE
XGBoost	Apache 2.0	https://github.com/dmlc/xgboost/blob/master/LICENSE
Pillow	HPND	https://github.com/python-pillow/Pillow/blob/main/LICENSE
seaborn	New BSD	https://github.com/mwaskom/seaborn/blob/master/LICENSE.md
SQLAlchemy	MIT	https://github.com/sqlalchemy/sqlalchemy/blob/main/LICENSE
spaCy	MIT	https://github.com/explosion/spaCy/blob/master/LICENSE
nltk	Apache 2.0	https://github.com/nltk/nltk/blob/develop/LICENSE.txt
boto3	Apache 2.0	https://github.com/boto/boto3/blob/develop/LICENSE
tabulate	MIT	https://github.com/astanin/python-tabulate/blob/master/LICENSE
future	MIT	https://github.com/PythonCharmers/python-future/blob/master/LICENSE.txt
jsonify	MIT	https://pypi.org/project/jsonify/0.5/#data
opencv-python	MIT	https://github.com/opencv/opencv-python/blob/4.x/LICENSE.txt
pyyaml	MIT	https://github.com/yaml/pyyaml/blob/master/LICENSE
Sentence Transformers	Apache 2.0	https://github.com/UKPLab/sentence-transformers/blob/master/LICENSE
wandb	MIT	https://github.com/wandb/wandb/blob/master/LICENSE
Deepspeed	Apache 2.0	https://github.com/microsoft/DeepSpeed/blob/master/LICENSE
CuPyCUDA12x	MIT	https://github.com/cupy/cupy/blob/main/LICENSE
Timm	Apache 2.0	https://github.com/huggingface/pytorch-image-models/blob/main/LICENSE
OmegaConf	New BSD	https://github.com/omry/omegaconf/blob/master/LICENSE
SciPy	New BSD	https://github.com/scipy/scipy/blob/main/LICENSE.txt
gradient	ISC	https://github.com/Paperspace/gradient-cli/blob/master/LICENSE.txt

Information about license types:

Apache 2.0: https://opensource.org/licenses/Apache-2.0
MIT: https://opensource.org/licenses/MIT
New BSD: https://opensource.org/licenses/BSD-3-Clause
PSF = Python Software Foundation: https://en.wikipedia.org/wiki/Python_Software_Foundation_License HPND = Historical Permission Notice and Disclaimer: https://opensource.org/licenses/HPND ISC: https://opensource.org/licenses/ISC

Open source software can be used for commercial purposes: https://opensource.org/docs/osd#fields-of-endeavor.

Software not included

Other software considered but not included.

The potential data science stack is far larger than any one person will use so we don't attempt to cover everything here.

Some generic categories of software not included:

• Non-data-science software
• Commercial software
• Software not licensed to be used on an available VM template
• Software only used in particular specialized data science subfields (although we assume our users probably want a GPU)

Category	Software	Why Not
Apache	Kafka, Parquet
Classifiers	libsvm	H2O contains SVM and GBM, save on installs
Collections	ELKI, GNU Octave, Weka, Mahout
Connectors	Academic Torrents
Dashboarding	panel, dash, voila, streamlit
Databases	MySQL, Hive, PostgreSQL, Prometheus, Neo4j, MongoDB, Cassandra, Redis	No particular infra to connect to databases
Deep Learning	Caffe, Caffe2, Theano, PaddlePaddle, Chainer, MXNet	PyTorch and TensorFlow are dominant, rest niche
Deployment	Dash, TFServing, R Shiny, Flask	Use Gradient Deployments
Distributed.	Horovod, OpenMPI	Use Gradient distributed
Feature store	Feast
IDEs	PyCharm, Spyder, RStudio
Interpretability	LIME/SHAP, Fairlearn, AI Fairness 360, InterpretML
Languages	R, SQL, Julia, C++, JavaScript, Python2, Scala	Python is dominant for data science
Monitoring	Grafana
NLP	GenSim
Notebooks	Jupyter, Zeppelin	JupyterLab includes Jupyter notebook
Orchestrators	Kubernetes	Use Gradient cluster
Partners	fast.ai	Could add if we want partner functionality
Pipelines	AirFlow, MLFlow, Intake, Kubeflow
Python libraries	statsmodels, pymc3, geopandas, Geopy, LIBSVM	Too many to attempt to cover
PyTorch extensions	Lightning
R packages	ggplot, tidyverse
Recommenders	TFRS, scikit-surprise
Scalable	Dask, Numba, Spark 1 or 2, Koalas, Hadoop
TensorFlow	TF 1.15, Recommenders, TensorBoard, TensorRT
Viz	Bokeh, Plotly, Holoviz (Datashader), Google FACETS, Excalidraw, GraphViz, ggplot2, d3.js