Georgian Io Hydra Versions Save

A cloud-agnostic ML Platform that will enable Data Scientists to run multiple experiments, perform hyper parameter optimization, evaluate results and serve models (batch/realtime) while still maintaining a uniform development UX across cloud environments

v0.3.9

3 years ago

We are now using public container repos to host each cloud's image of Hydra
Some bug fixes in Hydra dockerfile

v0.3.8

3 years ago

As part of release 0.3.8 we are releasing the following features :

Setup a scalable MLFlow setup using terraform (Read more here : https://github.com/georgian-io/hydra/wiki/Infrastructure-as-Code-(AWS))
Git is not no longer needed to work with Hydra
GPU related fixes for AWS cloud

v0.3.7

3 years ago

Fixing issues #38 and #29

v0.3.5

3 years ago

Fixed bug where local and fast_local modes where pointing to incorrect shell scripts
Fixed bug where options being passed to fast_local cloud mode were breaking training

v0.3.4

3 years ago

Jobs stuck in RUNNABLE state on AWS Batch
Reverting to pre gpu implementation

v0.3.3

3 years ago

ability to use gpu count for aws

v0.3.2

3 years ago

support for --cloud=="aws"

v0.3.1

3 years ago

Fixing issue https://github.com/georgianpartners/hydra/issues/21
Refactoring functions

v0.3.0

3 years ago

First public release
Improved documentation

0.1.0-alpha

3 years ago