名称 最后提交 最后更新
..
CMakeLists.txt add mkldnn prop_kind phase for inference-only case to pooling and activations (#14278)
attention_lstm_fuse_pass.cc code style fix
attention_lstm_fuse_pass.h fea/fuse attention lstm simplify.with fusion lstm.with sequnce expand (#13006)
conv_bias_mkldnn_fuse_pass.cc Change the origin VLOG level to 10 times
conv_bias_mkldnn_fuse_pass.h Conv+Bias: Support non-null bias
conv_bn_fuse_pass.cc Change the origin VLOG level to 10 times
conv_bn_fuse_pass.h conv bn fuse pass
conv_elementwise_add_mkldnn_fuse_pass.cc MKLDNN residual connections fuse pass: Maybe removed and boost::optional used where it makes sense
conv_elementwise_add_mkldnn_fuse_pass.h MKLDNN residual connections fuse pass: Maybe removed and boost::optional used where it makes sense
conv_elementwise_add_mkldnn_fuse_pass_tester.cc MKLDNN residual connections fuse pass: unit tests enabled and added
conv_relu_mkldnn_fuse_pass.cc Change the origin VLOG level to 10 times
conv_relu_mkldnn_fuse_pass.h add depthwise conv mkldnn pass
conv_relu_mkldnn_fuse_pass_tester.cc fix a few more tests
depthwise_conv_mkldnn_pass.cc Change the origin VLOG level to 10 times
depthwise_conv_mkldnn_pass.h add depthwise conv mkldnn pass
depthwise_conv_mkldnn_pass_tester.cc add depthwise conv mkldnn pass
embedding_fc_lstm_fuse_pass.cc - Yet another clarification to comment
embedding_fc_lstm_fuse_pass.h - Added initial pass for embedding-fc-lstm
fc_fuse_pass.cc add in_num_col_dims for fc
fc_fuse_pass.h fea/refine fuse (#13076)
fc_fuse_pass_tester.cc fix typo to pass the ci
fc_gru_fuse_pass.cc refactor ir pattern (#13304)
fc_gru_fuse_pass.h add fuse fc gru pass
fc_lstm_fuse_pass.cc refine peephole
fc_lstm_fuse_pass.h test/add text-classification test (#13081)
fuse_elewise_add_act_pass.cc Change the origin VLOG level to 10 times
fuse_elewise_add_act_pass.h Feature/op_fuse_pass (#12440)
fuse_pass_base.cc Add MKL-DNN placement pass (#13958)
fuse_pass_base.h Add MKL-DNN placement pass (#13958)
graph.cc Combine Inference Analysis with IR (#13914)
graph.h Combine Inference Analysis with IR (#13914)
graph_helper.cc Change the origin VLOG level to 10 times
graph_helper.h clean up
graph_helper_test.cc fix typo and warning in analyzer_resnet50_test
graph_pattern_detector.cc MKLDNN residual connections fuse pass:
graph_pattern_detector.h MKLDNN residual connections fuse pass:
graph_pattern_detector_tester.cc bugfix/fusion lstm (#13185)
graph_test.cc fix typo and warning in analyzer_resnet50_test
graph_to_program_pass.cc Combine Inference Analysis with IR (#13914)
graph_to_program_pass.h graph to program pass
graph_to_program_pass_test.cc fix graph_to_program_test to be platform-independent
graph_traits.cc Combine Inference Analysis with IR (#13914)
graph_traits.h Combine Inference Analysis with IR (#13914)
graph_viz_pass.cc Change the origin VLOG level to 10 times
graph_viz_pass.h refine/fc lstm fusion link (#13158)
infer_clean_graph_pass.cc bugfix/fusion lstm (#13185)
is_test_pass.cc add mkldnn prop_kind phase for inference-only case to pooling and activations (#14278)
is_test_pass.h add mkldnn prop_kind phase for inference-only case to pooling and activations (#14278)
is_test_pass_tester.cc add mkldnn prop_kind phase for inference-only case to pooling and activations (#14278)
mkldnn_placement_pass.cc Change the origin VLOG level to 10 times
mkldnn_placement_pass.h Add MKL-DNN placement pass (#13958)
multi_batch_merge_pass.cc Change the origin VLOG level to 10 times
multi_batch_merge_pass.h [1.1] [project] train imagenet using large batch size (#13766)
node.cc code style fix
node.h code style fix
node_test.cc add tests
pass.cc pass builder allow cutomize pass in python.
pass.h code style fix
pass_builder.cc pass builder allow cutomize pass in python.
pass_builder.h add some comments
pass_test.cc fix
seq_concat_fc_fuse_pass.cc Change the origin VLOG level to 10 times
seq_concat_fc_fuse_pass.h fea/fuse attention lstm simplify.with fusion lstm.with sequnce expand (#13006)
seqconv_eltadd_relu_fuse_pass.cc Change the origin VLOG level to 10 times
seqconv_eltadd_relu_fuse_pass.h add seqconv eltadd relu pass

项目简介

PArallel Distributed Deep LEarning: Machine Learning Framework from Industrial Practice (『飞桨』核心框架,深度学习&机器学习高性能单机、分布式训练和跨平台部署)

:rocket: Github 镜像仓库 :rocket:

源项目地址 :arrow_down: :arrow_down: :arrow_down:

https://github.com/paddlepaddle/paddle

deep-learningdistributed-trainingefficiencymachine-learningneural-networkpaddlepaddlepythonscalability

发行版本 60

PaddlePaddle 2.5.0 Release Note

全部发行版

贡献者 246

全部贡献者

开发语言

  • C++ 49.8 %
  • Python 41.0 %
  • Cuda 7.0 %
  • CMake 1.1 %
  • Shell 0.6 %
反馈
建议
客服 返回
顶部