๐Ÿฆ‰DVC Profile picture
Dec 20 โ€ข 6 tweets โ€ข 4 min read
๐Ÿ‘จ๐Ÿปโ€๐Ÿ’ป Ever tried deploying a model? and ended up entangled in scripts?

Hereโ€™s how MLEM lets us do โ€œsingle command deploymentsโ€ โšก๏ธ:

- Deployment Options
- Setting up the Environment
- Run the command
- Getting Predictions

@Iterativeai @DVCorg
#mlem #aws #deployment

๐Ÿงต[1/6]
๐Ÿš€ MLEM gives us a simple and powerful API to deploy to platforms such as:

โœ… @awscloud Sagemaker
โœ… Docker Container
โœ… Heroku
โœ… Kubernetes

๐Ÿงต[2/6]
๐ŸŒฑ Setting up the Environment

Setting up the environment varies for each case, but here let's take the example of Heroku.

We can either set โ€œHEROKU_API_KEYโ€ environment variable or use Heroku CLI to run โ€œheroku loginโ€.

๐Ÿงต[3/6]
๐Ÿƒโ€โ™‚๏ธ Run the command

Run โ€œmlem deployment run heroku app.mlem โ€”model <model> โ€”app_name <app name>โ€

Voila! All Done โœ…

๐Ÿงต[4/6]
๐Ÿ“Š Getting Predictions

Our model above is reachable by HTTP request, we can open the URL and see OpenAPI spec there, or send requests to get predictions.

We can also use built-in MLEM functionality for the same using โ€œmlem deployment apply app.mlem data.csv --jsonโ€

๐Ÿงต[5/6]
๐Ÿ˜Ž Go ahead, try it out yourself!

cml.dev

๐Ÿงต[6/6]

โ€ข โ€ข โ€ข

Missing some Tweet in this thread? You can try to force a refresh
ใ€€

Keep Current with ๐Ÿฆ‰DVC

๐Ÿฆ‰DVC Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @DVCorg

Dec 15
๐——๐—ฉ๐—–๐—Ÿ๐—ถ๐˜ƒ๐—ฒ + ๐˜†๐—ผ๐˜‚๐—ฟ ๐—ฐ๐—ผ๐—ฑ๐—ฒ = ๐—˜๐˜…๐—ฝ๐—ฒ๐—ฟ๐—ถ๐—บ๐—ฒ๐—ป๐˜ ๐˜๐—ฟ๐—ฎ๐—ฐ๐—ธ๐—ถ๐—ป๐—ด!

Produce an experiment tracker in CLI or @Code with DVCLive and a few lines added to your code or notebook

Version your ML experiments fast with no database or SaaS needed colab.research.google.com/drive/1VKEBdSgโ€ฆ
1/4
๐—›๐—ผ๐˜„ ๐—ฑ๐—ผ๐—ฒ๐˜€ ๐—ถ๐˜ ๐˜„๐—ผ๐—ฟ๐—ธ?

1 . Install DVC and DVCLive ๐š™๐š’๐š™ ๐š’๐š—๐šœ๐š๐šŠ๐š•๐š• --๐šž๐š™๐š๐š›๐šŠ๐š๐šŽ ๐š๐šŸ๐šŒ

2. Initialize your repo with ๐—š๐—ถ๐˜ and ๐——๐—ฉ๐—–

2/4
3 . In your code or notebook, enable DVC experiment tracking using DVCLive with ๐šœ๐šŠ๐šŸ๐šŽ_๐š๐šŸ๐šŒ_๐šŽ๐šก๐š™=๐šƒ๐š›๐šž๐šŽ

3/4
Read 4 tweets
Dec 14
โœจ Visualize your pipelines as beautiful graphs of connected stages using โ€œdvc dagโ€ command.

Here's different types of graphs you can explore ๐Ÿ—บ๏ธ

โœ… Full DAG
โžก๏ธ DAG of outputs
๐Ÿงœโ€โ™€๏ธ Mermaid flowchart
โšช๏ธ DOT format
๐Ÿ“ฃ Verbose

@Iterativeai @DVCorg
#data #pipeline #dag

๐Ÿงต[1/7]
โœ… Full DAG

โ€œdvc dag --fullโ€ shows the full DAG that the target stage belongs to, instead of showing only its ancestors.

๐Ÿงต[2/7]
โžก๏ธ DAG of outputs

โ€œdvc dag --outsโ€ shows a DAG of chained dependencies and outputs instead of the stages themselves.

๐Ÿงต[3/7]
Read 7 tweets
Dec 12
๐Ÿฆ‰ DVC is designed to improve upon past solutions to make the life of ML teams easier. Hereโ€™s how it differentiates from other related technologies:

๐Ÿฑ Git
๐Ÿ“ฆ Git-LFS (Large File Storage)
๐Ÿ—„ Git-annex

@Iterativeai @DVCorg
#mlOps #data #git #lfs

๐Ÿงต[1/5]
๐Ÿฑ Git vs DVC

DVC builds upon Git by the concept of data files โ€“ large files that should not be stored in a Git repository, but need to be tracked and versioned.

It leverages Git's features to enable managing different versions of data, data pipelines, and experiments.

๐Ÿงต[2/5]
๐Ÿ“ฆ Git-LFS vs DVC

DVC does not require special servers like Git-LFS demands. Any cloud storage like S3, Google Cloud Storage, or even an SSH server can be used as a remote storage.

No additional databases, servers, or infrastructure are required.

๐Ÿงต[3/5]
Read 5 tweets
Dec 3
๐Ÿฆ‰ Hereโ€™s a quick brief from Iterative Community๐Ÿ‘‡

๐Ÿช GitHub Universe
๐Ÿค Hacktoberfest Contributions
๐Ÿ“น Closure Session 2 event
โ™พ๏ธ CML at NeurIPS
๐Ÿ‡ฌ๐Ÿ‡ง MLOps Summit London
๐Ÿ‡บ๐Ÿ‡ธ ODSC West

@Iterativeai @DVCorg
#mlOps #data #pipeline #dag

๐Ÿงต[1/8]
๐Ÿช Dmitry at GitHub Universe

Dmitry Petrov (@FullStackML) spoke at #GitHubUniverse on โ€œML with Git: experiment tracking in Codespacesโ€ on 8th November.

He presented how to use the DVC extension for VS Code and Codespaces to streamline ML experimentation process.

๐Ÿงต[2/8]
๐Ÿค Hacktoberfest Contributions

Huge thanks to @francesco086, vvssttkk, and @DeepyamanDatta for their contributions!

We are sending personalized shirts to them.

Special thanks to @mertbozkirr for leading the Hacktoberfest charge!

๐Ÿงต[3/8]
Read 8 tweets
Nov 16
๐Ÿฆ‰Wanna get started with DVC on your own machine? Here's how you can get started๐Ÿ‘‡

๐ŸŽ MacOS
๐ŸชŸ Windows
๐Ÿ–ฅ๏ธ Linux
๐Ÿ Python Package
๐Ÿ‘จ๐Ÿปโ€๐Ÿ’ป VS Code Extension

@Iterativeai @DVCorg
#mac #windows #linux #code

๐Ÿงต [1/7]
๐ŸŽ MacOS

The recommended way to install dvc on Mac is using @MacHomebrew with command โ€œbrew install dvcโ€ โ˜•๏ธ

๐Ÿงต [2/7]
๐ŸชŸ Windows

If you are on Windows, we recommend installing Chocolatey. Then using the command โ€œchoco install dvcโ€ ๐Ÿซ

๐Ÿงต [3/7]
Read 7 tweets
Nov 7
๐Ÿฆ‰ Hereโ€™s some exciting news brief this month from Iterative Community ๐Ÿ‘‡

๐Ÿค October Meetup
๐ŸŽ‰ Hacktoberfest
๐Ÿฒ DVC-Hydra Integration
โ™พ๏ธ CML in Research
๐Ÿ“„ DVC Wikipedia Page

@Iterativeai @DVCorg
#mlOps #data #pipeline #dag

๐Ÿงต [1/7]
๐Ÿค October Meetup

@NadiaIIT presented her team's research on Collaboration Challenges in Building ML-Enabled Systems in the October Meetup. If you missed it, visit .

๐Ÿงต [2/7]
๐ŸŽ‰ Hacktoberfest

Contributions are pouring in, as Hacktoberfest proceeds. Find out all the issues with #hacktoberfest tags in our repos.

๐Ÿงต [3/7]
Read 7 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us on Twitter!

:(