Commit Graph

14 Commits (715d862868aca6ae4c865dc3db1cde6818e4ad1a)

Author SHA1 Message Date
Leo Chen c0163837a5
Fix compile problem when cuda_arch < 6000 (#29576)
4 years ago
Leo Chen 9f926eb720
Layernorm opt (#29522)
4 years ago
Leo Chen a040c055a5
fix layer_norm accuracy (#29434)
4 years ago
furnace 7584bb5096
Layer norm fp16 (#29169)
4 years ago
lijianshe02 9f83f0fe69
API/OP (group_norm, layer_norm, random_crop, unpool) error message enhancement (#24413)
5 years ago
mapingshuo 7d4002e06a
restrict block num of layer_norm_grad cuda block to 128 (#23878)
5 years ago
Pei Yang 0a51098a71
Add TRT support for BERT (#21135)
5 years ago
Yu Yang f57d706aa7 Use double to reduce
7 years ago
sneaxiy c50c537732 fix arithmetic error in backward kernel
7 years ago
sneaxiy 010883689c Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into refine_layer_norm
7 years ago
sneaxiy ad45d39222 refine layer_norm
7 years ago
qingqing01 24509f4af9 Fix the grammar in copyright. (#8403)
7 years ago
Yi Wang fc374821dd Correct #include path
7 years ago
Yi Wang 90648f336d Move file to fluid/; Edit CMakeLists.txt
7 years ago