Pilanci Research Group

Riemannian_Preconditioned_LoRA

source code for paper "Riemannian Preconditioned LoRA for Fine-Tuning Foundation Models"

Randomized-Geometric-Algebra-Methods-for-Convex-Neural-Networks

caldera

Compressing Large Language Models using Low Precision and Low Rank Decomposition

scnn

Scalable Convex Neural Networks

1d-convex-nn

nfda

scnn_experiments

Code to replicate experiments in https://arxiv.org/abs/2202.01331.

relu_optimal_sets

matrix-compressor

Implementation of LPLR algorithm for matrix compression

IHS-BIN

Threshold-networks

convex_autoencoder

Two-Layer ReLU Autoencoders via Convex Optimization based on the paper "Convex Neural Autoregressive Models: Towards Tractable, Expressive, and Theoretically-Backed Models for Sequential Forecasting and Generation" by B. Bartan, T. Ergen, M. Pilanci

Neural-recovery

poly_act_convex_NN

Adaptive-Newton-Sketch

convex_nn

adaptive_effdim_solver

Companion code for the paper "Fast Convex Quadratic Optimization Solvers with Adaptive Sketching-based Preconditioners"