AI as an API - Part 1 - Train an ML Model and turn it into an Rest API using Keras, FastAPI & NoSQL

Подписаться 348 тыс.

Просмотров 77 тыс.

50% 1

AI as an API - Part 1 - Turn a ML Trained Model into an Rest API using Keras, FastAPI & NoSQL
PART 2: • AI as an API - Part 2 ...
Learn to create & deploy a deep learning algorithm into a production REST API using Python, Keras, FastAPI, & NoSQL.
⦿ Build a Spam Classifier with Keras (post): kirr.co/u2tpsc
⦿ UCI Datasets: archive.ics.uci.edu/ml/datase...
⦿ Project Tutorials: cfe.sh/projects
⦿ Sign up for DataStax AstraDB: dtsx.io/3lC4cWy
⦿ Reference Code: github.com/codingforentrepren...
⦿ Production Code: github.com/codingforentrepren...
⦿ Subscribe: cfe.sh/youtube
0:00:00 - Welcome
00:00:32 - Demo
00:04:13 - Requirements
00:05:50 - Setup Project on VSCode
00:14:55 - Download Datasets
00:24:14 - Unzip Datasets
00:26:55 - Extract Review & Combine Dataset
00:42:46 - Convert Dataset into Vectors
00:58:26 - Split & Export Vectorized Dataset
01:09:43 - Training our Machine Learning Model
01:23:21 - Upload our Model to Object Storage
01:31:07 - Model Download Pipeline
01:49:10 - Base FastAPI App
01:56:05 - Load Keras Model into FastAPI
02:03:31 - Predictions with Keras & FastAPI
02:22:14 - Creating a Resuable AI Model Class
02:49:42 - AstraDB Keys, Pydantic Settings & Environment Variables
02:58:52 - Configure AstraDB Cluster & Session
03:08:20 - Cassandra Model in AstraDB
03:19:38 - Store inference data via Cassandra Model
03:23:28 - Paginate a Cassandra model as a FastAPI Streaming Response
03:38:47 - Ngrok, Pydantic Model, & Post Data
03:47:27 - Thank you

Опубликовано:

1 авг 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 67

@CodingEntrepreneurs 2 года назад

Part 2 is up! ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-nTdMjFcK3SM.html

@SanjeevKumar-nc2rt 2 года назад

I am going to watch whole video in 2 days. I am sure i am going to learn everything you taught in this video.. i will comments again my feedback after completion.😊😌

@amazing-graceolutomilayo5041 2 года назад

OH MY GOD!... I just saw this on my recommended. I have not seen any video yet. SUBBED!!!. I hope to learn more as I dig in

@abrarshahid3930 2 года назад

We are loving your FastAPI and Cassandra videos!! We want more......

@Sam98961 2 года назад

Such a great tutorial! Can't wait to start part 2 tomorrow. Thanks so much man.

@brandypiao5682 2 года назад

Only 42 mins in and I know you're a life saver

@MrAntivirus66 2 года назад

Sick tutorials recently.. thanks programing brother 🧬

@OnceARider 2 года назад

Nice content to pass the weekend. Thanks!

@robinsvantony 2 года назад

Started ML ... Thank you.

@agungokill 2 года назад

straightforward, nice!

@lambdamax 2 года назад

Thank you so much for doing this! There is hope in this sad world!

@MrTarekTV 2 года назад

Thank you for your valuable information, I appreciate your efforts, wish you the more success and the best

@BeattapeFactory 2 года назад

That sounds like a really cool idea. Thank you

@CodingEntrepreneurs 2 года назад

Thank you

@TheHunterhal 2 года назад

Thank you, haven't watched it yet but liked it, I will watch it this weekend and run the codes using Pytorch maybe image models (if possible) :).

@cedricvillani8502 2 года назад

Really? You’ll have to let us all know how that worked out, I’m kind of curious

@akashthoriya 2 года назад

Thank you sir

@prashantmorgaonkar3095 Год назад

Great work, lucid !

@alihusham1560 2 года назад

Man, after I learned from your tutorials I got a job on upwork thank you

@CodingEntrepreneurs 2 года назад

So cool! Nice work

@cedricvillani8502 2 года назад

Dude I saved 15% on my car insurance by switching to Geico thanks ;)

@harshprateek8045 2 года назад

Great content....

@tatg09 Год назад

Thanks!

@mihirpesswani5044 2 года назад

Please make a tutorial for OpenCV AI model with Django Rest API too.

@michas7993 2 года назад

Yeah, that’s right!

@onlyyou200548 2 года назад

FYI, model.predict is not thread-safe and hence, cannot be usef within fastAPI like that. A correct way is to use tensorflow serving

@krzysztofgalus3886 2 года назад

Great video 😎

@nepalcodetv6298 2 года назад

Gold Content

@frederickmai 2 года назад

Long time no see!

@siuu3155 2 года назад

Please make a tutorial on machine learning for beginners

@shashanksathish9362 2 года назад

This video taught me a lot, right from the best practices in Data Engineering/Data Science. However, I could not able to assimilate the use of AI as an API. I had built an AI project with front-end thus I could see the results directly, I would be glad to know the use case of the AI as an API like where this can be used. Apologies, if the question doesn't make sense.

@CodingEntrepreneurs 2 года назад

Great question! This api can be used in nearly *any* service. This means regardless what your front end or backend is, you can probably use this API. For example, if you have 4 different web applications, they can all use this API to provide predictions for the content. You can see this in action on my Try Django 3.2 series where we integrate Django with an AI microservice.

@shashanksathish9362 2 года назад

@@CodingEntrepreneurs Makes sense, perfect. Thanks!

@lllllMlllll 2 года назад

Ah UCL .. The university that nearly drove me crazy doing a master degree :)

@gabrielvl1 2 года назад

Awesome ... +1000(y)

@aloudiakite1217 2 года назад

Amazing tutorial, can you please make a tutorial on machine learning pipeline with datasets storage in Cassandra?

@CodingEntrepreneurs 2 года назад

Yes it's coming! Not in part 2 but in a future series!

@cedricvillani8502 2 года назад

IEEE dataport datasets are also fun and interesting, just FYI

@mehdismaeili3743 2 года назад

like like like.......😍

@shikunchen8523 2 года назад

Thank you for this wonderful video. I followed your tutorial, it worked fine. But for the function "export_inferences", the fastAPI docs gave me "Unrecognized response type; displaying content as text.", the dataset can't be showed. Do you have any idea why it happend?

@CodingEntrepreneurs 2 года назад

Perhaps turn your stream response into a string. Also thank you!

@brunoluan 2 года назад

What if I need the user to upload his dataset, choose the model and then get back the model trained? How I would do that?

@leonli5970 2 года назад

amazing video. MLops is silver bullet

@Tommybotham 10 месяцев назад

This is frustrating to follow as I'm encountering a: "WARNING: pip is configured with locations that require TLS/SSL, however the ssl module in Python is not available." Error. I'm using VS Code from Anaconda.

@markakritas8047 2 года назад

Thanks for very useful tutorial! Is there a free object storage available that will work also with boto3? Because as far as I understood all of three suggested solutions charge at least 5 $ / month, and it's unwanted to make an expense just for few files and testing purposes.

@CodingEntrepreneurs 2 года назад

Linode.com/cfe or do.co/cfe-sh For some free credits

@markakritas8047 2 года назад

@@CodingEntrepreneurs thanks a lot

@vipulsharma9094 2 года назад

The model is predicting everything as ham . i have followed each step in this video but the problem presists can anyone help me. The model itself at google colab is giving this result.

@HM_Milan 2 года назад

I make with this idea hehe but I didn't know how to

@isurumadusanka1590 2 года назад

Hi! Thanks for the huge tutorial. Eagerly waiting for the part 2. For some reason, part two link is not accessible (private link).

@CodingEntrepreneurs 2 года назад

Soon it will be published!!

@CodingEntrepreneurs 2 года назад

It’s here! ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-nTdMjFcK3SM.html

@isurumadusanka1590 2 года назад

@@CodingEntrepreneurs Thank you very much!

@akashthoriya 2 года назад

Sir, when we expect part 2 of this video? Description link says, private video. Please update sir

@CodingEntrepreneurs 2 года назад

It’s here!! ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-nTdMjFcK3SM.html

@akashthoriya 2 года назад

@@CodingEntrepreneurs Thank You!!!

@hazelnicolettemanners53 2 года назад

Why have u chosen fast api over django?

@dsdy1205 2 года назад

I read this as "Alas, an API"

@CodingEntrepreneurs 2 года назад

Haha!

@HugoCoolDude 2 года назад

okay... so why not pytorch, kedro and bentoml (and mlflow), that's waayyy easier than coding all stuff by hand, just watching this video was exhausting let alone coding all that by hand:p

@cedricvillani8502 2 года назад

To people that don’t have experience with higher level math like linear algebra, or algebraic combinatorics (Stochastic gradient dissent should be a concept that you know inside and out ) then this is not your video, if you’re looking for a quick fix go to hugging face there you can use sckitlearn and gardio

@vjukulkarni6057 2 года назад

Can we get the code ?

@CodingEntrepreneurs 2 года назад

Check the description

@FirstNameLastName-fv4eu 2 года назад

Everyone teaching ML but no one knows how to earn money from it

@nepalcodetv6298 2 года назад

you can use above api in comment system to approve and unapprove spam comments automatically based on percentage, make money selling feat or make softw like Disqus comment, duh ML, AI is future...

@FirstNameLastName-fv4eu 2 года назад

@@nepalcodetv6298 who will buy that ? Google is it ?