You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by GitBox <gi...@apache.org> on 2020/11/12 01:40:24 UTC
[GitHub] [incubator-mxnet] mseth10 commented on a change in pull request #19386: [1.x] Move block.optimize_for backend_opts to kwargs
mseth10 commented on a change in pull request #19386:
URL: https://github.com/apache/incubator-mxnet/pull/19386#discussion_r521709245
##########
File path: python/mxnet/gluon/block.py
##########
@@ -1087,19 +1087,23 @@ def optimize_for(self, x, *args, backend=None, backend_opts=None, clear=True, **
other inputs to model
backend : str
The name of backend, as registered in `SubgraphBackendRegistry`, default None
- backend_opts : dict of user-specified options to pass to the backend for partitioning, optional
- Passed on to `PrePartition` and `PostPartition` functions of `SubgraphProperty`
- clear : clears any previous optimizations
+ clear : bool, default False
+ Clears any previous optimizations
static_alloc : bool, default False
Statically allocate memory to improve speed. Memory usage may increase.
static_shape : bool, default False
Optimize for invariant input shapes between iterations. Must also
set static_alloc to True. Change of input shapes is still allowed
but slower.
+ **kwargs: The backend options, optional
+ Passed on to `PrePartition` and `PostPartition` functions of `SubgraphProperty`
"""
+ if len(kwargs) > 0:
+ self._backend_opts = kwargs
- # do hybrize API call
- self.hybridize(True, backend, backend_opts, clear, **kwargs)
+ if clear or not self._active:
Review comment:
why do we need this condition now?
##########
File path: python/mxnet/gluon/block.py
##########
@@ -1087,19 +1087,23 @@ def optimize_for(self, x, *args, backend=None, backend_opts=None, clear=True, **
other inputs to model
backend : str
The name of backend, as registered in `SubgraphBackendRegistry`, default None
- backend_opts : dict of user-specified options to pass to the backend for partitioning, optional
- Passed on to `PrePartition` and `PostPartition` functions of `SubgraphProperty`
- clear : clears any previous optimizations
+ clear : bool, default False
+ Clears any previous optimizations
static_alloc : bool, default False
Statically allocate memory to improve speed. Memory usage may increase.
static_shape : bool, default False
Optimize for invariant input shapes between iterations. Must also
set static_alloc to True. Change of input shapes is still allowed
but slower.
+ **kwargs: The backend options, optional
+ Passed on to `PrePartition` and `PostPartition` functions of `SubgraphProperty`
"""
+ if len(kwargs) > 0:
+ self._backend_opts = kwargs
- # do hybrize API call
- self.hybridize(True, backend, backend_opts, clear, **kwargs)
+ if clear or not self._active:
+ # do hybrize API call
Review comment:
nit: hybridize
##########
File path: python/mxnet/gluon/block.py
##########
@@ -1087,19 +1087,23 @@ def optimize_for(self, x, *args, backend=None, backend_opts=None, clear=True, **
other inputs to model
backend : str
The name of backend, as registered in `SubgraphBackendRegistry`, default None
- backend_opts : dict of user-specified options to pass to the backend for partitioning, optional
- Passed on to `PrePartition` and `PostPartition` functions of `SubgraphProperty`
- clear : clears any previous optimizations
+ clear : bool, default False
+ Clears any previous optimizations
static_alloc : bool, default False
Statically allocate memory to improve speed. Memory usage may increase.
static_shape : bool, default False
Optimize for invariant input shapes between iterations. Must also
set static_alloc to True. Change of input shapes is still allowed
but slower.
+ **kwargs: The backend options, optional
+ Passed on to `PrePartition` and `PostPartition` functions of `SubgraphProperty`
"""
+ if len(kwargs) > 0:
+ self._backend_opts = kwargs
- # do hybrize API call
- self.hybridize(True, backend, backend_opts, clear, **kwargs)
+ if clear or not self._active:
+ # do hybrize API call
+ self.hybridize(True, backend, clear, static_alloc=static_alloc, static_shape=static_shape)
Review comment:
we might have to rebase this PR, as we recently added another parameter `partition_if_dynamic` to both hybridize and optimize_for
##########
File path: python/mxnet/gluon/block.py
##########
@@ -1137,7 +1141,11 @@ def register_child(self, block, name=None):
super(HybridBlock, self).register_child(block, name)
self._clear_cached_op()
- def hybridize(self, active=True, backend=None, backend_opts=None, clear=True, **kwargs):
+ def hybridize(self, active=True, backend=None, clear=True,
Review comment:
we still need kwargs for hybridize, right?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org