Deploy ML models with FastAPI, Docker, and Heroku | Tutorial

In this Tutorial you'll learn how to deploy Machine Learning models with FastAPI, Docker, and Heroku.
Code: github.com/AssemblyAI-Example...
Resources:
fastapi.tiangolo.com/deployme...
devcenter.heroku.com/categori...
github.com/basil-b2s/Language...
Get your Free Token for AssemblyAI Speech-To-Text API 👇www.assemblyai.com/?...
▬▬▬▬▬▬▬▬▬▬▬▬ CONNECT ▬▬▬▬▬▬▬▬▬▬▬▬
🖥️ Website: www.assemblyai.com
🐦 Twitter: / assemblyai
🦾 Discord: / discord
▶️ Subscribe: ua-cam.com/users/AssemblyAI?...
🔥 We're hiring! Check our open roles: www.assemblyai.com/careers
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
Timeline:
00:00 Introduction
01:08 Model Development
05:00 FastAPI app
08:44 Dockerize the app
14:14 Deployment to Heroku
17:46 Final Testing
#MachineLearning #FastAPI

[@davegordon6233]

Best explanation I found in 3 days! Thank you.

[@wuttipatsricharoensuk5072]

Thank you so much for sharing this excellent tutorial. I followed the steps, and I was able to successfully deploy my first model. Your clear explanations and step-by-step guidance made the process really smooth and enjoyable.

[@ary_21]

heroku doesnt have a free tier anymore , you have a paid subscription?

[@wuttipatsricharoensuk5072]

@@ary_21 I haven't subscribed to any paid plan on Heroku. I just deleted the app once I finished deploying it. Since it avoids any unnecessary costs while still taking advantage of learning and experimentation.

[@ary_21]

@@wuttipatsricharoensuk5072got it, actually
I have interviews in next 2 months so looking for.a solution for free deployment
Being. A fresher i cant afford aws etc
If you know any do let me know

[@MrAmgadHasan]

@@ary_21 render has a free tier that offers 512GB of RAM and 0.1 CPU.
You can deploy fastapi (and even plotly dash and streamlit) there but your models need to be able to run with these limited resources.

[@EGlobalKnowledge]

Thank you very much, was able to deploy the whole thing

[@boisiul]

really great tutorial, thanks man!

[@checher100]

absolutely great! Thanks for that

[@caxupin6666]

Thanks for the video!

[@ajeetsbpr]

Thanks for explaining this.

[@codecrush_666]

thanks a lot for sharing, excellent tutorial!

[@MrShihab100]

Thanks for this professional tutorial :)

[@muhammadumarsotvoldiev8768]

thank you. This video helped me a lot

[@TheSkogemann]

Thank you! 👏

[@currentactivities6527]

thank you for sharing ❤

[@theiyobosa]

Amazing!

[@TAP7a]

The filling in of the requirements.txt part is why all python users should use envs by default, all the time :)

[@ivan07valdes]

What a good video!

[@aminatashnezhad]

great job.

[@ElskiezVazcez]

thanks mate

[@jonahturner2969]

Hey, thanks for the tutorial! We've noticed that our production FastAPI services all suffer from memory leaks, any suggestions/experience with that?

[@jamm9848]

Thank you for the video. Will deployment work for large files for example an ml model of up to 2 Gb in size?

[@manhpham5772]

Great tutorial! But could you tell me how to deal with docker-compose and Heroku, I have frontend and backend containers...

[@FiringTap]

genius!

[@hpatelecom1]

Briliant Patric

[@mendezluis3346]

Hello, I have an error, when I deploy the model on Heroku, and then open the app, display a r15 and r14 error, what should i do? I try the model with tensorflow-cpu

[@mansamusa559]

Thanks! Does anyone know how to easily deploy an ML Model which requires GPU ? (Like ine tuned stable diffusion model)

[@caiyu538]

Great

[@noobjok3r640]

can you help me like i want to a image classification ml model to deploy but in ur video its string version

[@ollieq5259]

Hey, at 14:52, you're using iTerm2 as your terminal in MacOS right? I also use it. But how did you make it pretty, specifically getting that 'arrow' shape over the text, and making part of the text blue? :)

[@Vicitoriachi]

He's using the zsh (Z-shell) unix shell, instead of bash ;).

[@ahb5819]

Heroku no longer has an free tier, any other recommendations?
Btw great video 😊

[@ashishmehra5143]

Vercel is a free option but I have not seen a success after deployment. I am able to run the application locally but when not after the deployment.

[@saifulislam-iv9px]

Möge Gott dich segnen!

[@user-vh3xp4rt5r]

Can you give link to google colab notebook?

[@alice.access5]

for bigger models, doesn't this cost a lot to host and run per month?

[@smiley3239]

other than heroku, when can we deploy it?

[@FiringTap]

If i have my idea can i deploy public ?

[@googlecolab7388]

like your smile

[@cesaravalos7591]

2:00 Portugeese 🦆🦆🦆🤣🤣🤣

[@deepjyotibaishya7576]

How to train in deep

[@adia9791]

Shouldnt you use ohe?

[@markcuello5]

HELP

[@thisiskartik]

Its a great video but I need to give a thumbs down because you never made this video a beginner friendly, Docker is really new to me as well has deploying it on heroku, it would have been great if you just a 1min introduction to what they are or where can we lear, like a prerequisite in the beginning of the video.

[@smilebig3884]

too basic.

[@GorTer-Balyants]

Why you talk like that

[@GorTer-Balyants]

why you talk like that