You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
Paddle/paddle/fluid/operators/optimizers
Gradie d79fdc3d62
lamb_op_xpu;test=kunlun (#31012)
5 years ago
..
CMakeLists.txt
adadelta_op.cc
adadelta_op.cu
adadelta_op.h
adagrad_op.cc
adagrad_op.cu
adagrad_op.h
adam_op.cc Support pure fp16 training for AMP API. (#29544) 5 years ago
adam_op.cu Support pure fp16 training for AMP API. (#29544) 5 years ago
adam_op.h Support pure fp16 training for AMP API. (#29544) 5 years ago
adam_op_xpu.cc fix header file paths of gflags, commit 3, test=develop (#30273) 5 years ago
adamax_op.cc
adamax_op.cu
adamax_op.h
decayed_adagrad_op.cc
decayed_adagrad_op.cu
decayed_adagrad_op.h
dgc_momentum_op.cc
dgc_momentum_op.cu
dgc_momentum_op.h
dpsgd_op.cc
dpsgd_op.h
ftrl_op.cc enhance error message, test=develop (#30220) 5 years ago
ftrl_op.cu
ftrl_op.h
lamb_op.cc Modify the calculation logic of LambOptimizer (#29313) 5 years ago
lamb_op.cu
lamb_op.h Modify the calculation logic of LambOptimizer (#29313) 5 years ago
lamb_op_xpu.cc lamb_op_xpu;test=kunlun (#31012) 5 years ago
lars_momentum_op.cc
lars_momentum_op.cu
lars_momentum_op.h
momentum_op.cc Support pure fp16 training for AMP API. (#29544) 5 years ago
momentum_op.cu
momentum_op.h Support pure fp16 training for AMP API. (#29544) 5 years ago
momentum_op_xpu.cc
proximal_adagrad_op.cc
proximal_adagrad_op.cu
proximal_adagrad_op.h
proximal_gd_op.cc
proximal_gd_op.cu
proximal_gd_op.h
rmsprop_op.cc
rmsprop_op.cu
rmsprop_op.h
rmsprop_op_xpu.cc add rmsprop_op_xpu test=kunlun (#30493) 5 years ago
sgd_op.cc
sgd_op.cu
sgd_op.h
sgd_op_xpu.cc
unity_build_rule.cmake