A game theoretic approach to explain the output of any machine learning model.

QLoRA: Efficient Finetuning of Quantized LLMs

PyTorch code and models for the DINOv2 self-supervised learning method.

PyHessian is a Pytorch library for second-order based analysis and training of Neural Networks

FeatureBooster: Boosting Feature Descriptors with a Lightweight Neural Network (CVPR 2023)

Mutual Information Neural Estimation in Pytorch

Official Implementation of paper "A Tale of Two Features: Stable Diffusion Complements DINO for Zero-Shot Semantic Correspondence"

Holds code for our CVPR'23 tutorial: All Things ViTs: Understanding and Interpreting Attention in Vision.

Framework code with wandb, checkpointing, logging, configs, experimental protocols. Useful for fine-tuning models or training from scratch, and testing them on a variety of datasets (transfer learn…

MegaDetector is an AI model that helps conservation folks spend less time doing boring things with camera trap images.

(ICLR 2023) Official PyTorch implementation of "What Do Self-Supervised Vision Transformers Learn?"

Code for "On the Spectral Bias of Neural Networks", to appear in ICML 2019 (Long Beach, CA).

A-SDF: Learning Disentangled Signed Distance Functions for Articulated Shape Representation (ICCV 2021)

"Understanding Dataset Difficulty with V-Usable Information" (ICML 2022, outstanding paper)

👋 Code for : "CRAFT: Concept Recursive Activation FacTorization for Explainability" (CVPR 2023)

Nonnegative spatial factorization for multivariate count data

Deep Feature Factorization

This is an official implementation for [ICLR'24] INTR: Interpretable Transformer for Fine-grained Image Classification.