Manan Shah's Projects
Paper Implementation for "Parameter-Efficient Transfer Learning for NLP"
bigint is a C++ library which can handle Very very Big Integers. It can calculate factorial of 1000000... it can go any big. It may be useful in Competitive Coding and Scientific Calculations which deals with very very large Integers. It can also be used in Decryption process. It has many inbuilt functions which can be very useful.
Unofficial Implementation to CDUL: CLIP-Driven Unsupervised Learning for Multi-Label Image Classification [ICCV'23]
A chrome extension for customising the codeforces website
Cross Domain Generalised Category Discovery
My personal blog
Solutions to CS307: Systems Practicum
A telegram chatbot that answers basic queries
Deep learning operations reinvented (for pytorch, tensorflow, jax and others)
Repo for solutions of Hacking Google Maps!
Codebase for the 1st Place Solution to the NCVPRIPG23 Challenge on Writer Verification
A pytorch like deep learning library built from scratch, with several common deep learning architectures, which can perform reverse mode automatic differentiation and has custom numpy like backends.
🗂 Split folders with files (i.e. images) into training, validation and test (dataset) folders by symlinks!
A unified framework for 3D content generation.
Torchmetrics - Machine learning metrics for distributed, scalable PyTorch applications.