Moving normalization core to impl - softmax (FX Converter Refactor [12/N]) <Target: converter_reorg_elementwise> #1909
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
github-actions
bot
added
component: api [Python]
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
There was a problem hiding this comment.
There are some changes that do not conform to Python style guidelines:
--- py/torch_tensorrt/fx/converters/aten_ops_converters.py 2023-05-10 06:06:19.166184 +0000
+++ py/torch_tensorrt/fx/converters/aten_ops_converters.py 2023-05-10 06:06:48.425711 +0000
@@ -359,17 +359,11 @@
target: Target,
args: Tuple[Argument, ...],
kwargs: Dict[str, Argument],
name: str,
) -> Union[TRTTensor, Sequence[TRTTensor]]:
- return softmax(
- network,
- target,
- SourceIR.ATEN,
- name,
- kwargs["input"],
- kwargs["dim"])
+ return softmax(network, target, SourceIR.ATEN, name, kwargs["input"], kwargs["dim"])
@tensorrt_converter(torch.ops.aten.cat.default)
def aten_ops_cat(
network: TRTNetwork,
--- py/torch_tensorrt/fx/converters/impl/normalization/__init__.py 2023-05-10 06:06:19.166184 +0000
+++ py/torch_tensorrt/fx/converters/impl/normalization/__init__.py 2023-05-10 06:06:48.631521 +0000
@@ -1 +1 @@
-from .ops import *
\ No newline at end of file
+from .ops import *
--- py/torch_tensorrt/fx/converters/impl/normalization/ops.py 2023-05-10 06:06:19.166184 +0000
+++ py/torch_tensorrt/fx/converters/impl/normalization/ops.py 2023-05-10 06:06:48.739746 +0000
@@ -12,20 +12,21 @@
from torch_tensorrt.fx.types import TRTNetwork, TRTTensor
from torch_tensorrt.fx.converters.converter_utils import (
SourceIR,
set_layer_name,
- get_positive_dim
+ get_positive_dim,
)
+
def softmax(
network: TRTNetwork,
target: Target,
source_ir: Optional[SourceIR],
name: str,
input: TRTTensor,
- dim: Optional[Any] = None
+ dim: Optional[Any] = None,
) -> Union[TRTTensor, Sequence[TRTTensor]]:
input_ranks = len(input.shape) + (1 if network.has_implicit_batch_dimension else 0) # type: ignore[union-attr]
if not isinstance(input, TRTTensor):
raise RuntimeError(
@@ -53,6 +54,5 @@
layer = network.add_softmax(input)
layer.axes = 1 << dim
set_layer_name(layer, target, name)
return layer.get_output(0)
-
--- py/torch_tensorrt/fx/converters/acc_ops_converters.py 2023-05-10 06:06:19.166184 +0000
+++ py/torch_tensorrt/fx/converters/acc_ops_converters.py 2023-05-10 06:06:51.503546 +0000
@@ -857,18 +857,11 @@
target: Target,
args: Tuple[Argument, ...],
kwargs: Dict[str, Argument],
name: str,
) -> Union[TRTTensor, Sequence[TRTTensor]]:
- return softmax(
- network,
- target,
- SourceIR.ACC,
- name,
- kwargs["input"],
- kwargs["dim"]
- )
+ return softmax(network, target, SourceIR.ACC, name, kwargs["input"], kwargs["dim"])
@tensorrt_converter(acc_ops.tile)
def acc_ops_tile(
network: TRTNetwork,
--- py/torch_tensorrt/fx/test/converters/aten_op/test_softmax_aten.py 2023-05-10 06:06:19.174185 +0000
+++ py/torch_tensorrt/fx/test/converters/aten_op/test_softmax_aten.py 2023-05-10 06:06:51.919309 +0000
@@ -39,6 +39,6 @@
TestModule(), input_specs, expected_ops={torch.ops.aten._softmax.default}
)
if __name__ == "__main__":
- run_tests()
\ No newline at end of file
+ run_tests()
apbose
requested review from
frank-wei,
narendasan and
wushirong
and removed request for
yinghai
narendasan
changed the title
apbose
changed the title
gs-olive
force-pushed
the
converter_reorg_elementwise
branch
from
0d4e6d7 to
9611d67
Compare
gs-olive
force-pushed
the
converter_reorg_normalization_softmax
branch
from
2769db2 to
2f23743
Compare
apbose
force-pushed
the
converter_reorg_elementwise
branch
from
9611d67 to
93846ed
Compare
Signed-off-by: Naren Dasan <[email protected]> new file: ../converters/impl/unary/base.py
narendasan
force-pushed
the
converter_reorg_elementwise
branch
from
93846ed to
83986cd
Compare
… <Target: converter_reorg_elementwise> (#1905)
narendasan
force-pushed
the
converter_reorg_normalization_softmax
branch
from
2f23743 to
f423be1
Compare
softmax linting error fix
narendasan
force-pushed
the
converter_reorg_normalization_softmax
branch
from
f423be1 to
75b1a2a
Compare
wushirong
approved these changes
Jun 13, 2023
apbose
force-pushed
the
converter_reorg_elementwise
branch
from
546f975 to
c8a9559
Compare
apbose
force-pushed
the
converter_reorg_elementwise
branch
2 times, most recently
from
18e503b to
2caac76
Compare
|
Closed with #2070 merged with main |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.