Build A PyTorch Style Transfer Web App With Streamlit

In this tutorial we build an interactive deep learning app with Streamlit and PyTorch to apply style transfer. This tutorial should demonstrate how easy interactive web applications can be build with Streamlit. Streamlit lets you create beautiful apps for your machine learning projects with simple Python scripts.

~~~~~~~~~~~~~~ GREAT PLUGINS FOR YOUR CODE EDITOR ~~~~~~~~~~~~~~

🚀🚀 SUPPORT ME ON PATREON 🚀🚀

🚀🚀 JOIN MY NEWSLETTER 🚀🚀

If you enjoyed this video, please subscribe to the channel!

Code:

Style Transfer code:

You can find me here:

#Python #PyTorch

----------------------------------------------------------------------------------------------------------
* This is a sponsored link. By clicking on it you will not have any additional costs, instead you will support me and my project. Thank you so much for the support! 🙏

[orjihvy]

this is a cool project. and how you separated out training and loading to leverage caching is really helpful.

[starlite]

Thanks for the video. I didn't understand what training a new model would do. You give it a dataset and an image that would be your style-image. What happens during the training?

[CodeWithTomi]

I love this guy.... Keep it up patrick

[dataaholic]

Love your content man ! Thanks for the video

[sultanakalsekar]

First Comment, .,. Amazing Getting Highly Inspired by U

[pratikkhandelwal]

Thanks for the video Patrick. Great content as usual. Just wanted to ask you - I am trying to load my custom content images and I have corresponding style image too. But after executing the training script, I am not getting very clear stylized output. I have tried varying content and style weights along with the epoch part. But results are not close to the pre-trained models. Any thoughts on what hyper-parameters should I choose? Thanks again :)

[tuanrobot]

how about the source image is uploaded manually by user? is it possible to make web app with streamlit like that?

[FindMultiBagger]

Really nice tutorial :-

Can you make seperate tutorial on how to achieve catching on different models without streamlite in python.

Example :

In fast ai we always need to load model before predict, if it's wrote in pure python script.

✅
So, How do we load any model Once and predict anything ?
With same concept ?


Please help !

Thanks 👌🙂

[deepakdeepu]

hi, this is Deepak, i really enjoy your video and learn stuff for that u really thank you alot. just one request please consider this can your please can a video on Deepfack images . please consider my request please

[muhammadbilalsiddiqui]

Thank you so much. Please let me know how can I stylize my own picture other thn save in model??