Andrey Malevich ec51f887bf Create only one instance of SigridTransform in DPerExample. ... Summary: DPer example have been creating multiple copies of the transform config in net defition till this moment, that resulted in the fact that I've hit the limit of ProtoBuf (64MB) for a certain Task requests (especially visible because of the ValidationPipeline that I was adding). After this diff we're going to store SigridTransforms in one instance per machine for training (or 1 instance per reading). Difference in sizes of the plans for some simple SparseNN model ~30 MB (even including the fact that second model have validation plan as well). TODO: Do similar logic for NNPreProc as well (it's also pretty large). Reviewed By: dzhulgakov Differential Revision: D4441441 fbshipit-source-id: 4452dd86a4dc49b2c7f5b7642f443aed5720b047		2017-01-22 19:29:16 -08:00
..
__init__.py	fbsync. TODO: check if build files need update.	2016-11-15 00:00:46 -08:00
batch_lr_loss.py	fbsync at f5a877	2016-11-18 15:41:06 -08:00
concat.py	fbsync at f5a877	2016-11-18 15:41:06 -08:00
dot_product.py	implement sparse nn using layers	2016-11-29 15:18:38 -08:00
expand_dims.py	implement sparse nn using layers	2016-11-29 15:18:38 -08:00
fc.py	fbsync at f5a877	2016-11-18 15:41:06 -08:00
layers.py	Create only one instance of SigridTransform in DPerExample.	2017-01-22 19:29:16 -08:00
simple_operator_layers.py	Fix random issues with some of the layers getting missing from registry.	2017-01-10 15:14:31 -08:00
sparse_lookup.py	implement user-only metadata for input_record	2016-12-15 12:01:29 -08:00
sparse_to_dense.py	implement user-only metadata for input_record	2016-12-15 12:01:29 -08:00
split.py	implement sparse nn using layers	2016-11-29 15:18:38 -08:00
tags.py	fbsync. TODO: check if build files need update.	2016-11-15 00:00:46 -08:00