Expanding linear RNN state-transition matrix eigenvalues to include negatives improves state-tracking tasks and language modeling without added training or inference costs.

An implementation of the paper "Parallelizing Linear Transformers with the Delta Rule over Sequence Length"

code released for our ICML 2020 paper "Do We Really Need to Access the Source Data? Source Hypothesis Transfer for Unsupervised Domain Adaptation"

Triton implement of bi-directional (non-causal) linear attention

🚀 Efficient implementations for emerging model architectures

Uni-MoE: Lychee's Large Multimodal Model Family.

[ICCV 2023] Official repository of FLatten Transformer

Pytorch library for fast transformer implementations

[ICLR 2023 Spotlight] Code release for "Dirichlet-based Uncertainty Calibration for Active Domain Adaptation"

Pytorch implementations of density estimation algorithms: BNAF, Glow, MAF, RealNVP, planar flows

Official implementation of paper "DenseHybrid: Hybrid Anomaly Detection for Dense Open-set Recognition"

Neural Spline Flow, RealNVP, Autoregressive Flow, 1x1Conv in PyTorch.

Code for "Implicit Normalizing Flows" (ICLR 2021 spotlight)

source code for ICLR'22 paper "VOS: Learning What You Don’t Know by Virtual Outlier Synthesis"

Official implementation for "Out-of-Distribution Detection in Long-Tailed Recognition with Calibrated Outlier Class Learning" (AAAI'24)

Flowcon: Out-of-Distribution Detection using Flow-Based Contrastive Learning [In Progress]

Density-Softmax: Efficient Test-time Model for Uncertainty Estimation and Robustness under Distribution Shifts (ICML 2024).

We present a new method for long-tailed out-of-distribution detection

The Official Implementation of the ICCV-2021 Paper: Semantically Coherent Out-of-Distribution Detection.

Official implementation for "Partial and Asymmetric Contrastive Learning for Out-of-Distribution Detection in Long-Tailed Recognition" (ICML'22 Long Presentation)

Implementation of Real NVP in PyTorch

This is reimplementation of "Simple and Principled Uncertainty Estimation with Deterministic Deep Learning via Distance Awareness" in Pytorch.

High-quality implementations of standard and SOTA methods on a variety of tasks.

Implemented CURE algorithm from robustness via curvature regularization and vice versa

[ICML 2024] Official code for Uncertainty Estimation by Density Aware Evidential Deep Learning

ICML 2024 Spotlight. This code is the official implementation for the paper, "Memorization Through the Lens of Curvature of Loss Function Around Samples"

Offical Implementation for "Curvature Clues: Decoding Deep Learning Privacy with Input Loss Curvature", NeurIPS '24 Spotlight