Python Callbacks: Which Not Everybody Knows
There is a very interesting thing I learnt which is to build our own Learner class. So Learner class is going to be a very general purpose training loop, which we can get it to do anything that we want it to do. And we gonna be creating things called callbacks to to make that happen.
Convolutions: In-Depth Understanding
This notebook provides an in-depth explanation of essential concepts in convolutional neural networks (CNNs), including filters, strides, and receptive fields. It explores the process of finding edges (top, left, diagonal) in an image using convolutional kernels and demonstrates how to apply these kernels to image regions. Additionally, the notebook covers the step-by-step construction of a CNN model from scratch, offering a comprehensive understanding of CNN architecture and operations. Whether you’re new to CNNs or looking to deepen your understanding, this notebook serves as a valuable resource for learning and experimentation.
Minibatch Training With Validation
This notebook demonstrates the implementation of a basic neural network training loop using custom parameters and optimization functions. Built from scratch, it includes the setup of a dataset and data loader, as well as validation procedures. The focus is on showcasing the foundational concepts of training a neural network using minimalai, a framework developed entirely from scratch in Python. This notebook provides a clear and concise guide to understanding the fundamental components of neural network training.
Animal Classifier: Identifying 90 Different Animals with Fastai
Explore my animal classifier project, where I trained a deep learning model to identify 90 different animals using Fastai. Check out the notebook to dive into the details of the model training process and see the classifier in action. If you find it useful, feel free to give it a try and provide feedback!
Car Classifier: Error rate 0.10 with fast.ai in 7 Epochs
Explore my car classifier project, where I trained a deep learning model to identify 196 different cars using Fastai. Check out the notebook to dive into the details of the model training process and see the classifier in action. If you find it useful, feel free to give it a try and provide feedback!