Borna Ahmadzadeh

attorch. A subset of PyTorch's neural network modules, written in Python using OpenAI's Triton.

trap. Autoregressive transformers in APL

neural-network-cuda. Neural network from scratch in CUDA/C++

flaim. Flax Image Models - State-of-the-art pre-trained vision backbones for Flax.

attention-in-vision. PyTorch implementation of popular attention mechanisms in vision

aplearn. APL machine learning library

minbpe-hs. Byte-level byte pair encoding (BPE) in Haskell

vit-pytorch. PyTorch implementation of the vision transformer

rusten. A minimal, didactic implementation of tensors in Rust.

llaf. LLMs in Futhark

image-2. Image super-resolution & enhancement with perceptual loss and the U-Net architecture

tello-face-tracker. A face tracker for the Ryze Tello drone

efficientnet-pytorch. PyTorch implementation of EfficientNet

simmim-pytorch. PyTorch implementation of SimMIM

ada. Source-to-source autodiff for APL

simsiam-pytorch. PyTorch implementation of SimSiam

boom. Feature embeddings for Open Library books

open-clip-jax. Implementation of OpenAI's CLIP in JAX/Flax

programming-challenges. Competitive programming problems & solutions