Meditron is a suite of open-source medical Large Language Models (LLMs).

Accelerating your LLM training to full speed! Made with ❤️ by ServiceNow Research

DISCO is a code-free and installation-free browser platform that allows any non-technical user to collaboratively train machine learning models without sharing any private data.

Codebase for ICML submission "DOGE: Domain Reweighting with Generalization Estimation"

Language Identification with Support for More Than 2000 Labels -- EMNLP 2023

Code for NeurIPS 2024 Spotlight: "Scaling Laws and Compute-Optimal Training Beyond Fixed Training Durations"

trying to make WebGPU a bit easier to use

Freeing data processing from scripting madness by providing a set of platform-agnostic customizable pipeline processing blocks.

Tensor computation with WebGPU acceleration

distributed trainer for LLMs

Landmark Attention: Random-Access Infinite Context Length for Transformers

GPT in TensorFlow.js

StableLM: Stability AI Language Models

nanoGPT-like codebase for LLM training

The simplest, fastest repository for training/finetuning medium-sized GPTs.

Practical low-rank gradient compression for distributed optimization: https://arxiv.org/abs/1905.13727

ColTraIn HBFP Training Emulator

Robust Cross-lingual Embeddings from Parallel Sentences

Decentralized Privacy-Preserving Proximity Tracing -- Documents

Example code and applications for machine learning on Graphcore IPUs

Stochastic Gradient Push for Distributed Deep Learning

Introduction to PyTorch Workshop at the AMLD 2019

Open Challenge - Automatic Training for Deep Learning

Unsupervised Scalable Representation Learning for Multivariate Time Series: Experiments

Code and data for the WSDM '19 paper "Crosslingual Document Embedding as Reduced-Rank Ridge Regression (Cr5)"

Learn how to design, develop, deploy and iterate on production-grade ML applications.