NVARC solution to ARC-AGI-2

Fast analytical implementation of batch eigen-decomposition for 3x3 symmetric matrices with Pytorch. > 250x faster than regular Pytorch implementation of batch eigen-decomposition on GPU.

Evolutionary algorithm discovery using Claude Code

Testing baseline LLMs performance across various models

Unofficial implementation of Tiny Recursive Model (TRM), improvement to HRM from Sapient AI, by Alexia Jolicoeur-Martineau

A collection of LogitsProcessors to customize and enhance LLM behavior for specific tasks.

velocity to Seismic CUDA speed up

Simple RL training for reasoning

Our solution for the arc challenge 2024

NeMo Retriever Library is a scalable, performance-oriented document content and metadata extraction microservice. NeMo Retriever extraction uses specialized NVIDIA NIM microservices to find, contex…

A missing piece of the Python multitask (both threads and processes) API: An extension that supports stateful worker pools & size-aware iterators.

LLM KV cache compression made easy

Create an AI capable of solving reasoning tasks it has never seen before

Quick implementation of nGPT, learning entirely on the hypersphere, from NvidiaAI

Implementation of Rotary Embeddings, from the Roformer paper, in Pytorch

A tool to configure, launch and manage your machine learning experiments.

The AdEMAMix Optimizer: Better, Faster, Older.

Curated list of datasets and tools for post-training.

Closed-form Continuous-time Neural Networks

Code of the NVIDIA winning solution to the 2nd OGB-LSC at the NeurIPS 2022 challenge with dataset PCQM4Mv2

Slides for my presentation

Adan: Adaptive Nesterov Momentum Algorithm for Faster Optimizing Deep Models

For optimization algorithm research and development.

Official PyTorch implementation of the Fishr regularization for out-of-distribution generalization

Fast, flexible and easy to use probabilistic modelling in Python.