268 - How to deploy your trained machine learning model into a local web application?

Code generated in the video can be downloaded from here:
github.com/bns...
This video explains the process of using Flask to deploy your scikit-learn (or other) trained model into a web application.

[@LiveLifeWithLove]

Dr Sreeni what you and other developers making videos on Machine Learning should know that you are doing a fantastic job & truly using this medium for spreading education. God bless you sir!

[@DigitalSreeni]

Thank you so much 🙂 Knowledge is useless unless we use it to enlighten others.

[@merlinzbeard1]

Thank you so much for this video sir! It was easy to follow and you extremely easy to learn. I had a quick question though. If I were running a classification model (Multi class), how do I change the code to show which class is predicted for the input?

[@richardmilian9959]

Great tutorial master, thank you so much for your teaching.
I'll do it and practice

[@vidyasvidhyalaya]

sir can you please upload a seperate video for image classification model into a local web application? take an example video from your cahnnel "vgg16 based feature extraction and xgboost classification video"...sir.......please

[@servatechtips]

Sir, thanks for you very much, Sir can we get the lesson about YOLOv5 in object detection and classification.

[@easydatascience2508]

Excellent explanation, thank you !

[@pratikdash.4003]

Sir, You r Ossum... Thanks for understanding what a Student actually want's from his teacher.

[@inhibited44]

Is it possible to load one image into --> prediction_RF = RF_model.predict(input_img_features)[0] -- after the confusion matrix and predict what it is? my image is (256, 256) and the error says "ValueError: X has 256 features, but RandomForestClassifier is expecting 32768 features as input.
" I see that 8 x8 * 512 in the feature_extractor is 32768.

[@sakifschannel]

Dear sir, I've followed your process. But I'm facing an error, "numpy.ndarray has no attribute predict".
Please help 😞

[@AstroPatel]

I recall some time ago learning that pickle had the issue of being machine-specific - is this no longer a problem? Could you load the pickled modeled on a different machine than the deployment server?

[@ntaimorebotsane6970]

Good day sir... I am having dimension problems when resizing dimensions. I am tying to deployed an image classification model. Any advice???

[@nayanjittalukdar823]

Sir, can we do it for semantic segmentation using deep learning models ?

[@DigitalSreeni]

Yes. You need the user to provide an image input and display the segmented output. I will show the image input part in the upcoming tutorials.

[@nayanjittalukdar823]

@@DigitalSreeni please make a video as soon as possible.. it will be very helpful . Thank you for replying 🙂

[@Am-ug9np]

pickle load gives me file not found error. Any advice? I have the same file structure as you and my load statement is the same.

[@GMAtiqurRahaman]

Thanks a lot!! Your tutorials (new or old) at some points in the time are benefitting me!!

[@manishakumari-lz7pw]

Well explained, compact and precise. Thank you.

[@aashishmanandhar5188]

throughout the video i was wondering how would a motorbike make anyone healthy lol

[@MbaliNene-kf6wk]

Thank you so much Sir ..this is really great! your patience set you apart

[@asemalashory911]

That's amazing 😍
Doctor can you tell me what is name of program that you using to write code

[@chitsogdrelwa8179]

Excellent and understandable explanation! Very kind to explain even the basics to cover all range of learners. Thank you

[@princeequalizer3715]

You are such a great teacher. Thank you sir!

[@Hellojsm8823]

Wow, that's what i'm looking for.

[@f.b.i.5562]

Thank you sir, but this only works for numerical data type inputs.
What if my input has different data types (String, float, int). How can I make my endpoint at the backend handle it?

[@ElectromecanicaIndustrial]

useful video, thanks

[@muruganandamganesan4952]

Sreeni: Thanks for the video with very clear explanation! You are very knowledgeable. Please upload more in the future.

[@ShubhamKumar-zo6cu]

Lots Of Love From India Sir :)

[@lazysultan2223]

Where do we place the backend?

[@radityafijarpradana1484]

Thanks a lot for the tutorial. This is very helpful.

[@umairsabir6686]

Hi,
Appreciated work !
Will u also make a video for model deployment in edge devices related to tf serving and docker ?

[@bacs_aditisingri]

Thank u so much , can you please tell me if this code works for audio classification..?

[@juansebastian6284]

I think it should. But you will need a different type of input. In the video he shows a simple model that uses two numeric inputs, so simple text boxes are enough. However, for audio classification you have to upload an audio file and that you must handle and send it to the backend (Flask app).

[@behnoudshafizadeh3281]

thanks for your explanation, Mr. Sreeni

[@ArgonautaDigital]

Thanks nice video! Recently we moved to FastAPI witch has better performance and auto-generated swagger doc ^^

[@davidkudabo2440]

This is nice man.... Clean and straightforward.

[@zeeshankhanyousafzai5229]

You are the best for rewinding the things.
Love you

[@themusicfactory263]

Very helpful. Thank you.

[@hariharaan.s6921]

Thank u sir

[@ahaditab6364]

excellent work

[@mahmodaldahol]

Thank you very much

[@nirmalaravind3184]

fire

[@kunal_chand]

Insanely helpful

[@user-wr4yl7tx3w]

Superb explanation

[@불루이보스]

Thank you so much!

[@DigitalSreeni]

You're welcome!

[@mustafakemalpasha983]

awesome video

[@DigitalSreeni]

Glad you enjoyed it

[@mustafakemalpasha983]

@@DigitalSreeni Tensorflow or scikit-learn for beginners? ( For production)

[@DigitalSreeni]

Doesn't matter, whatever solves the problem. If speed is important on an edge device, look into neural network approach where models can be optimized for edge devices.

[@mustafakemalpasha983]

​@@DigitalSreeni Perfect! thanks for the answer man. I will look into it.