Paddle

Commit Graph

Author	SHA1	Message	Date
Shang Zhizhou	b9e76a0103	detect tensorRT plugin fp16 in runtime (#27933 ) * remove -DSUPPORTS_CUDA_FP16 in cuda.cmake * comile with cuda9 * add some unittest * notest;test=coverage * add unittest for trt plugin swish && split * update ernie unittest * fix some error message * remove repeated judgement of CUDA version in mbEltwiseLayerNormOpConverter * fix comile errror when CUDA_ARCH_NAME < Pascal" * fix comile error * update unittest timeout * compile with cuda9 * update error msg * fix code style * add some comments * add define IF_CUDA_ARCH_SUPPORT_FP16 * rename IF_CUDA_ARCH_SUPPORT_FP16 to CUDA_ARCH_FP16_SUPPORTED	6 years ago
Shang Zhizhou	47fdc60ecc	Optimize slice trt plugin (#26970 ) * optimize slice TRT plugin This patch removes unnecessary barrier for data transfer of needed offset, so data transfer can be overlap with GPU kernel execution. This patch also fixes incorrect name of slice plugin. That is, replaces "layernorm" with "slice" test=develop * add serialize/deserialize to slice plugin * add static shape slice trt plugin * fix slice trt op convertor dynamic shape bug * fix format by clang-format * fix pylint format error * fix problems commented by peiyang Co-authored-by: Ryan Jeng <rjeng@nvidia.com>	6 years ago
Zhaolong Xing	133f1fc123	[Eernie TRT]: add slice op and add emb eltwise layernorm fp16 support (#23723 ) * refine ernie trt dynamic shape support 1. add slice op converter 2. add emb eltwise layernorm fp16 support test=develop * fix dynamic shape test ut test=develop * fix comments. test=develop * fix comments test=develop	6 years ago

Author

SHA1

Message

Date

Shang Zhizhou

b9e76a0103

detect tensorRT plugin fp16 in runtime (#27933 )

* remove -DSUPPORTS_CUDA_FP16 in cuda.cmake

* comile with cuda9

* add some unittest

* notest;test=coverage

* add unittest for trt plugin swish && split

* update ernie unittest

* fix some error message

* remove repeated judgement of CUDA version in mbEltwiseLayerNormOpConverter

* fix comile errror when CUDA_ARCH_NAME < Pascal"

* fix comile error

* update unittest timeout

* compile with cuda9

* update error msg

* fix code style

* add some comments

* add define IF_CUDA_ARCH_SUPPORT_FP16

* rename IF_CUDA_ARCH_SUPPORT_FP16 to CUDA_ARCH_FP16_SUPPORTED

Shang Zhizhou

47fdc60ecc

Optimize slice trt plugin (#26970 )

* optimize slice TRT plugin

This patch removes unnecessary barrier for data transfer of needed offset,
so data transfer can be overlap with GPU kernel execution.

This patch also fixes incorrect name of slice plugin. That is, replaces
"layernorm" with "slice"

test=develop

* add serialize/deserialize to slice plugin

* add static shape slice trt plugin

* fix slice trt op convertor dynamic shape bug

* fix format by clang-format

* fix pylint format error

* fix problems commented by peiyang

Co-authored-by: Ryan Jeng <rjeng@nvidia.com>

Zhaolong Xing

133f1fc123

[Eernie TRT]: add slice op and add emb eltwise layernorm fp16 support (#23723 )

* refine ernie trt dynamic shape support
1. add slice op converter
2. add emb eltwise layernorm fp16 support
test=develop

* fix dynamic shape test ut
test=develop

* fix comments.
test=develop

* fix comments
test=develop

3 Commits (cced930b61ba246dffec68bbe09bd9e22a142d64)