mikailkhona's repositories
stepwise_inference_icml24
This repository contains code for ICML 2024: Towards an Understanding of Stepwise Inference in Transformers: A Synthetic Graph Navigation Model
activation_weight_quant
Repo to test various methods and speedups for activation and weight quantization in pytorch
columnformers
A Transformer-inspired model of the brain
CUDA_kernels
Here is a collection of CUDA kernels for small scale simulations. Intended to be used as a guide to learn CUDA programming
Double-Descent
Repo to study double descent in linear regression. This is to build intuition.
mikailkhona
Config files for my GitHub profile.
NPEET
Non-parametric Entropy Estimation Toolbox
pythia
The hub for EleutherAI's work on interpretability and learning dynamics
spaghetti
SPAtial GrapHs: nETworks, Topology, & Inference
sparsegpt
Code for the ICML 2023 paper "SparseGPT: Massive Language Models Can Be Accurately Pruned in One-Shot".