Image Classification With Streamlit| Deep Learning WebApp|

Deploy a deep learning keras model onto a web app using Streamlit.
We start by training a image classifier in tensorflow ans keras and then saving the model in google collab.
We then create a web application using streamlit and finally host the web server on the internet using ngrok application.
You will be able to deploy the ml web app directly from google collab.(Deep learning model deployment directly from google colab)

Recommended books for getting better at ML Web Apps and Flask:

[NachiketaHebbar]

I have added the source code in the description. Since some issues have been reported by subscribers in the deployment, i have also added an alternative deployment code which does not need ngrok authentication code. Check it out and let me know in the comments which code worked for you.

[sairamadithya]

you're rocking!!! cant express the amount of happiness i have right now

[rikiriki]

Thank you so much for your video. I am new to the programming language. Your course is understandable.

[anmoldubey]

This was awesome!!
Thank you so much...it really helped a lot!!!

[RenitaRianti]

hii.. your video is very interesting, but I found a problem when i try it. my localhost wont connect, do you know how to fix it?

[anamikamaurya]

nicely explained.. god bless you with more and more enthu and knowledge

[plabmadeeasy]

I had to pip install pyngrok and then import ngrok from there before running the ngrok auth token

[pratik]

Some of you might get different results if ur doing this on ur transfer learning models, this lil modified code will help u in getting the better and accurate results in prediction. change the import and predict code to this

from skimage import transform
size = (224, 224)
image = ImageOps.fit(image_data, size, Image.ANTIALIAS)
image = np.asarray(image)

np_image = transform.resize(image, (224, 224, 3))
np_image = np.expand_dims(np_image, axis=0)
img = np_image

[suharinic]

hii Nachiketa, the complete code has been executed successfully but at the end i am not able to open the website, can you please help me with that

[sarfarazahmed]

Hi Nachiketa, do you have any tutorial related to deploying a deep learning model which identifies the speaker.

[thiruranjith]

failed to complete tunnel connection what to do

[jacobturney]

Hi! Thank you so much for the walkthrough. I am having trouble with the line:
!nohup streamlit run app.py &

I keep getting different errors. One is a syntax error and the other is "OSError: Background processes not supported.
"

I used python in Spyder to build my model and have been using Jupyter Notebooks for everything else. I am using a Windows machine. Any advice?

[johnlouieiligan]

This is a nice video.!!

What if i want to use my images locally and not using from tensorflow. how you should do it? Thanks

[vishwadesai]

I am getting an error like /bin/bash: ngrok: command not found while executing authentication cell

[soumyadrip]

Great video...did not know about ngrok...thanks.
Creating a video on using Docker in Heroku to deploy would be a great video...and helpful for us

[impanaimpu]

ERROR: google-colab 1.0.0 has requirement ipykernel~=4.10, but you'll have ipykernel 5.5.5 which is incompatible.
/bin/bash: streamlit: command not found
what to do?

[vanillasky]

So when I run my code it is showing an error like "file object has no attribute predict" how can I fix this?

[Mayglie]

not able to run !ngrok authtoken in colab. Please help!!
getting this message>> /bin/bash: grok: command not found.

[mihretdesta]

Good Boy, but how about print accuracy...

[anandruparelia]

Hello! Thanks for the video! It helped me alot!
BTW, With which video editor we can shoot the video like yours?
Displaying screen + The presenter in the small frame?