Build a WhatsApp Media Cleaner with Deep Learning - Kartik Godawat, DeepKlarity

Livestream & Recording: https://youtu.be/o1Di7T2vdYM

Please use this thread to ask questions during and after the webinar.

Summary
In this webinar, we’ll create a simple image classification model using the FastAI deep learning library to classify WhatsApp images & videos into categories like photos, memes, receipts, greetings etc. This model can help clean up photos in bulk when your device is running low on storage. This talk is beginner friendly, and no prior knowledge of machine learning is required to attend. ML concepts will be explained at a high level using code & examples.

About The Speaker
Kartik Godawat is the co-founder of DeepKlarity and a Kaggle Competition Expert where he is currently ranked in top 500. He is also an active blogger on medium and is keenly interested in computer vision, nlp and generative modeling. Kartik likes trying out the latest trends in the ML community. He also has an active interest in applying ML to drone piloting systems and brain-computer interfaces.

there is noise of echo

Image download original:

Image download (FastAI):

WhatsApp cleaner model:

What if I have a single folder with images and a CSV/Excel/TXT file with labels. Can we still use ImageDataBunch?

pls explain how the CNN is working with images here

@VishalMaurya We will cover this (Convolutional Neural Networks) in a lot of detail in one of our upcoming webinars in the “Deep Learning with PyTorch” series.

Sure Aakash :slightly_smiling_face: