Ray RLlib

Ray Train

Ray Datasets

RayDP

Retail

Gaming

Government

Financial

Ray Serve

Anyscale

Ray Core

Healthcare

Technology

Ray Tune

News

User Story

Engineering

Culture

Kai Fricke

Figure 4. Training times for number of GPUs

Figure-4

Faster stable diffusion model fine-tuning with Ray AIR

Faster stable diffusion fine-tuning with Ray AIR

Jiao Dong

Hao Zhang

Lianmin Zheng

Jun Gong

Jules S. Damji

Phi Nguyen

This is part 2 of our generative AI blog series. Here, we cover how Ray empowers large language models (LLM) frameworks such as Alpa. 

Figure 5: Technical integration layered stack for LLM


Figure-5

High performance LLM training at 1000 GPU scale with Alpa and Ray

Training 175B Parameter Language Models at 1000 GPU scale with Alpa and Ray

Antoni Baum

Eric Liang

Eric Liang headshot

Richard Liaw

Richard Liaw headshot

This is part 1 of our generative AI blog series. In this blog, we talk about how to use Ray to deploy common generative model workloads in production. An upcoming blog will deep dive into why projects like Alpa are using Ray to scale large models.

layered-stack

Ray solves Generative AI and LLM Infrastructure Challenges

How Ray solves common production challenges for generative AI infrastructure

Anyscale Ray Team

How companies use Ray to scale many Python jobs at massive scale. Scaling many parallel Python jobs can come in many different forms. You can use Ray to train one million jobs in record time, scale many simulations or train the number of experiments that are part of a hyperparameter job.


summit-stories-scaling-parallel-python-jobs

How companies are scaling parallel python jobs for many models training, simulation or hyperparameter tuning.

Ray Summit Series - Scaling Parallel Python Jobs

Ray Summit Stories - ML Workloads (1)

How companies are scaling their ML workloads and driving efficiency and productivity

Ray Summit 2022 stories - ML Platforms

Clarence Ng

Figure 2: OOM monitor high-level architecture and data flow

Figure 2

Automatic and optimistic memory scheduling for ML workloads in Ray

Uday Kanwar

Deb Daipayan

Blue River Tractor

Blue River Technology is revolutionizing agriculture by applying advanced computer vision and robotics to the problem of crop production using the Anyscale fully-managed Ray platform

Blue River Technology Developers Iterate 2.5X Faster with the Anyscale Fully-Managed Ray Platform

Cade Daniel

Zhe Zhang

The Ray 2.3 release features exciting improvements across the Ray ecosystem. In this blog post, we will highlight new features, performance enhancements, and support for new platforms. 

main_release_image

Announcing Ray 2.3: performance improvements, new features and new platforms

Large Language Models

Ray Summit 2022 Stories - Large Language Models

 Max Mergenthaler

Together, Ray, Fugue and Nixtla provide a powerful open-source solution for organizations looking to perform forecasting and anomaly detection at scale.

Nixtla and Ray

Forecasting at Scale With Nixtla and Ray | Anyscale

Forecasting at Scale

Using Anyscale with Weights & Biases simplifies MLOps with a built-in integration that allows for scalability, reproducibility and consistency.

Weights&Biasesflow

Simplify ML Development Cycle with Anyscale and Weights & Biases

Simplify your ML Development Cycle with Anyscale and Weights & Biases

Frank Sifei Luan, UC Berkeley

frank

Exoshuffle is now the most cost-efficient way to sort 100TB of data on the public cloud, using only $0.97 per terabyte. Learn how this is possible with Ray.

CloudSortBenchmark 01

Ray Is the World’s Most Cost-Efficient Sorting System at $1/TB

Ray breaks the $1/TB barrier as the world’s most cost-efficient sorting system

More companies are needing to train & deploy many small ML models, often hundreds or thousands. Learn how Ray and Anyscale make that possible.

Many_model_image

Training 1 Million ML Models in Record Time | Anyscale

Training One Million Machine Learning Models in Record Time with Ray

This section is used to order the "Types" and "Tags" that show up for filters on the Blog Index

Types

Products / Libraries

From the creators of Ray, Anyscale is a framework for building machine learning applications at any scale originating from the UC Berkeley RISELab.

Blog

Scalable AI and Python

Featured Posts and News

Training 175B Parameter Language Models at 1000 GPU scale with Alpa and Ray

How Ray solves common production challenges for generative AI infrastructure

Ray breaks the $1/TB barrier as the world’s most cost-efficient sorting system

No posts found.