[
https://issues.apache.org/jira/browse/IGNITE-8795?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Anton Dmitriev updated IGNITE-8795:
---
Description: As described in the [design
document|https://docs.google.com/document/d/1jROIahK1rc7bSgOvhJhfpMqIGvht_IE8zn5NAt6x8ks/edit?usp=sharing],
Distributed TensorFlow is based on TensorFlow cluster concept. It's a set of
TensorFlow processes started among the cluster and available througth the gRPC
interfaces. It's assumed that these processes contain heavy operations that
requires data to be stored locally on the nodes where the processes running.
Apache Ignite admits the data to be moved from one node to another as result of
node failure of rebalancing. As result the TensorFlow cluster should be changed
dynamically as well as TensorFlow Cache (follow-the-data strategy). (was: As
described in the design document
(https://docs.google.com/document/d/1jROIahK1rc7bSgOvhJhfpMqIGvht_IE8zn5NAt6x8ks/edit?usp=sharing)
Distributed TensorFlow is based on TensorFlow cluster concept. It's a set of
TensorFlow processes started among the cluster and available througth the gRPC
interfaces. It's assumed that these processes contain heavy operations that
requires data to be stored locally on the nodes where the processes running.
Apache Ignite admits the data to be moved from one node to another as result of
node failure of rebalancing. As result the TensorFlow cluster should be changed
dynamically as well as TensorFlow Cache (follow-the-data strategy).)
> Add ability to start and maintain TensorFlow cluster on top of Apache Ignite
>
>
> Key: IGNITE-8795
> URL: https://issues.apache.org/jira/browse/IGNITE-8795
> Project: Ignite
> Issue Type: New Feature
> Components: ml
>Reporter: Yury Babak
>Assignee: Anton Dmitriev
>Priority: Major
> Fix For: 2.6
>
>
> As described in the [design
> document|https://docs.google.com/document/d/1jROIahK1rc7bSgOvhJhfpMqIGvht_IE8zn5NAt6x8ks/edit?usp=sharing],
> Distributed TensorFlow is based on TensorFlow cluster concept. It's a set of
> TensorFlow processes started among the cluster and available througth the
> gRPC interfaces. It's assumed that these processes contain heavy operations
> that requires data to be stored locally on the nodes where the processes
> running. Apache Ignite admits the data to be moved from one node to another
> as result of node failure of rebalancing. As result the TensorFlow cluster
> should be changed dynamically as well as TensorFlow Cache (follow-the-data
> strategy).
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
