MLearning.ai
Published in

MLearning.ai

The secret of deploying GPT-3 app

How to deploy GPT-3 app? We will deploy State-of-the-Art NLP model in matter of minutes.

Photo by Robert Gramner on Unsplash

Free tier account

We will use two APIs to deploy GPT-3. Both are available via free tier. Credit card is not required. The source code is hosted in your Github repository. That is all.

The GPT-3 is available via OpenAI API. It will require registration. The free tier will grant you 18 USD of free credit. There is no need for a credit card. The registration process is self-explanatory otherwise.

We have now account created. Log in to the API. Then click the “View API keys”-section. We are able to generate new secret keys in this page. Be careful not to share these secret keys.

Deploy

We generate the app via Streamlit. Streamlit has free tier. We will use it to deploy our app. The source code of this app is hosted on your Github repository.

A sample Github repository is available here. The repository includes the app.py file. It is coded in Python. I share an example app here. It will help you get started.

The Github repository includes Requirements.txt too. See an example here. It defines the packages required by our Streamlit app.

The Streamlit will need access to your Github repository. The OpenAI secret keys are inserted to the app advance settings.

We can now deploy the app.

Register your app

The app is now available. Congratulations! It may seem too simple, but you have an app available to showcase to anybody.

The deployment of GPT-3 comes with a responsibility. Such large language models are subject to misuse. Toxic speech is one such application.

GPT-3 apps are validated by OpenAI to prevent misuse. The deployment documentation and review process is explained here.

Be sure to submit your app for the validation. This way your API access is not revoked.

--

--

--

Data Scientists must think like an artist when finding a solution when creating a piece of code. ⚪️ Artists enjoy working on interesting problems, even if there is no obvious answer ⚪️ linktr.ee/mlearning 🔵 Follow to join our 18K+ Unique DAILY Readers 🟠

Recommended from Medium

Deposit and Withdraw PancakeSwap on Wollito.

Building a Virtual City (Round Two)

Taste all the features of Hyperledger Fabric v1.4

Tech is not all about coding!

Azure Web App- Spring Boot Backend Service

Self Closure Memory leak fixes using Weak and Unowned

What should we build next?

CSS3 in 10 days — Day 3

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Teemu Maatta

Teemu Maatta

Machine Learning Engineer. Top writer in Natural Language Processing (NLP). Multimodal learning. Artificial General Intelligence (AGI). DALL·E 2. GPT-3.

More from Medium

Reducing offensive content in NLP

Analyses Of GPT-3 Paper

University of Amsterdam & Meta AI Propose a Roadmap Toward Interactive Language Modelling Based on…

Learning how GPT-3 instruct models were (most likely) trained