[IGNITE-7456] Fix wrong batch logic in distributed MLP training.

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.4
• Fix Version/s: 2.4
• Component/s: ml
• Assignee: Artem Malykh
• Reporter: Artem Malykh
• Created: 17/Jan/18 14:12
• Updated: 23/Jan/18 10:25
• Resolved: 23/Jan/18 10:25

Description

Batch for training is created outside of training loop, therefore in each local step we work with the same batch.

Issue Links

• links to: GitHub Pull Request #3398
• links to: lic & javadoc
• links to: TC run