Managing The Lifetime and Registration of TensorRT Plugins

Computing Higher-Order Derivatives Using Automatic Differentiation

Derivation and Extension of The Classical Optimal Brain Surgeon Algorithm

First Time Serving as ICML Area Chair

TensorRT Implicit Weight Quantization Caveats and Tricks

Jacobian Matrix, De Novo Chain Rule Expression, Jacobian-Vector Product, Vector-Jacobian Product

Sharing Key and Value Tensors for a Group of Query Tensors to Mitigate Transformer Attention Layer Performance Bottleneck

Performance Bottleneck for Serving Transformer Models

Same Performance as Group-Wise Weight-Only Quantization But with Better Accuracy

First Time Serving as NeurIPS Area Chair

Scene Representation and Differentiable Rendering with Neural Radiance Fields

Fine-Tuning and Pruning of Large Language Models Using Low-Rank Adaptation