40 lines
1.5 KiB
ReStructuredText
40 lines
1.5 KiB
ReStructuredText
|
:orphan:
|
||
|
|
|
||
|
|
.. _grid_cloud_run_expert:
|
||
|
|
|
||
|
|
###########################
|
||
|
|
Train on the cloud (expert)
|
||
|
|
###########################
|
||
|
|
**Audience**: Corporate or academic users who want to run on their Company or University private cloud account.
|
||
|
|
|
||
|
|
----
|
||
|
|
|
||
|
|
********************************
|
||
|
|
Run on your own cloud or cluster
|
||
|
|
********************************
|
||
|
|
If you have access to a corporate or academic cluster, you can simply submit jobs that run Lightning models and lightning will automatically work.
|
||
|
|
Please refer to :doc:`Clusters guide <cluster>` for more information.
|
||
|
|
|
||
|
|
----
|
||
|
|
|
||
|
|
***********************************
|
||
|
|
Run on your own cloud (hassle free)
|
||
|
|
***********************************
|
||
|
|
Cluster training can get complicated once you start doing multi-node training, fault-tolerant training or sweeps.
|
||
|
|
If you'd prefer to not deal with any of the hassles of running on your own cloud environments, lightning-grid enables University and Enterprise customers to run on the cloud with their own credentials or even onprem.
|
||
|
|
|
||
|
|
These are some of the benefits of running via lightning-grid:
|
||
|
|
|
||
|
|
- create datasets optimized for scale
|
||
|
|
- fully configurable on-prem deployment
|
||
|
|
- SOC-2 compliance (in-progress) (ETA Q3 2022)
|
||
|
|
- micro cost optimizations everywhere (which add up)
|
||
|
|
- built-in fault tolerance
|
||
|
|
- enabled collaboration for teams and enterprises
|
||
|
|
|
||
|
|
Contact our sales support engineering team so we can help you set up Grid with your own cloud credentials.
|
||
|
|
|
||
|
|
Email us to connect with your own cloud account:
|
||
|
|
|
||
|
|
`<onprem@pytorchlightning.ai>`_.
|