Experiment, data, and analysis code for "Text embedding models yield detailed conceptual knowledge maps derived from short multiple-choice quizzes" by Paxton C. Fitzpatrick, Andrew C. Heusser, and Jeremy R. Manning

Attention is all you need implementation

This project implements a hierarchical deep learning approach for recognizing group activities in videos. The core idea is to model both individual person actions and group-level activities using temporal dynamics captured by LSTM networks.

gpt-2 reincarnate architecture for reddit posts.

A modern PyTorch implementation of hierarchical deep temporal models for group activity recognition

llama-2 from scratch — a clean, educational pytorch implementation of the llama-2 transformer architecture. features grouped query attention (gqa), rotary position embeddings (rope), kv caching, and swiglu activation. run easily on google colab and experiment with your own datasets.

Implementation of Tversky Neural Networks, from Doumbouya et al. (2025), in PyTorch .

program synthesis with neuro-symbolic differentiable interpreters

Personal study and implementation of the "Evolution of Heuristics" paper using LLMs to generate heuristic algorithms.

A modern implementation of ECCV 2018 paper: "Hierarchical Relational Networks for Group Activity Recognition and Retrieval".

A multi-label text classifier based on PLM-ICD paper

Naive implementations of deep reinforcement learning algorithms

attention is all you need — pytorch implementation of the original transformer architecture for english to nepali neural machine translation (nmt), achieving around 27 bleu score.

Implementation of VAEs (Variational Autoencoders) with PyTorch.

Paper Reproduction: "DropBlock: A regularization method for convolutional networks"

A simple implementation based on the "Attention is All You Need" paper, using GPT-2 for text generation.

Baseline paper implementations of deep learning models and architectures in PyTorch.

Implementation of a Denoising Diffusion Probabilistic Models (DDPM) using PyTorch based on the original paper.

Mono Implementations' Archive

Implementation of the paper "Improved DeepFake Detection Using Whisper Features"