This post will contain resources I have used and found helpful in my learning journey of PyTorch and Deep Learning. I will update this post as I discover more resources.
Coding Videos
PyTorch for Deep Learning & Machine Learning : Beginner PyTorch course taught by Daniel Bourke. It covers the basics of PyTorch, neural networks, and applications in computer vision. Code-first approach with hands-on examples in Google Colab and course textbook, Learn PyTorch for Deep Learning: Zero to Mastery book.
MNIST classifier from numpy scratch: George Hotz builds a MNIST digit classifier from scratch using only numpy. Great for understanding the fundamentals of neural networks and backpropagation.
Deep Learning Research
Theory
Vision Transformers - A research paper on Vision Transformers (ViT) for image recognition. Walkthrough implementation by Daniel Bourke available.
Multiscale Vision Transformers - A research paper on Multiscale Vision Transformers (MViT) for video and image recognition. The model is available on GitHub.
MViTv2: Improved Multiscale Vision Transformers for Classification and Detection - An improved version of MViT
Cardiology
EchoPrime: A Multi-Video View-Informed Vision-Language Model for Comprehensive Echocardiography Interpretation - A research paper on a vision-language model for echocardiography interpretation. The code is implemented in PyTorch and available on GitHub.
EchoCLIP: Vision–language foundation model for echocardiogram interpretation - A research paper on a vision–language foundation model for echocardiography. The code is implemented in PyTorch and available on GitHub.
Merlin: a computed tomography vision–language foundation model and dataset - A research paper on a 3D VLM that learns from volumetric CT scans, electronic health record data and radiology reports. The code is implemented in PyTorch and available on GitHub.
PanEcho: Complete AI-enabled echocardiography interpretation with multi-task deep learning - A research paper on a multi-task deep learning model for echocardiography interpretation. The code is implemented in PyTorch and available on GitHub.
EchoNext-Mini: A Dataset and Baseline AI Model for Detecting Structural Heart Disease from Electrocardiograms - A research paper on a dataset and baseline AI model for detecting structural heart disease from electrocardiograms. The code is implemented in PyTorch and available on GitHub.
Opthalmology
Deep learning applications in opthamology - Talk by Aaron Y. Lee that covers the beginnings of the application of deep learning in the field of ophthalmology and vision science.
Transforming Healthcare with AI: Lessons from Ophthalmology - Talk by Pearse Keane on how AI is transforming healthcare, with a focus on ophthalmology.
Electronic Health Record (EHR)
- InfEHR: Clinical phenotype resolution through deep geometric learning on electronic health records - A research paper on using deep geometric learning to analyze electronic health records for clinical phenotype resolution. The code is implemented in PyTorch and available on GitHub.