名称 最后提交 最后更新
..
jit Make CreateProgramDesc more robust (#31543)
tests [HybridParallel] fix port reuse when create multi group (#31876)
CMakeLists.txt flush denormal in the tracer op, test=develop (#32350)
README.md fix sample code in paddle/fluid/imperative/README.md (#22141)
all_reduce.cc [ROCM] update fluid imperative for rocm (part1), test=develop (#31017)
all_reduce.h [ROCM] update fluid imperative for rocm (part1), test=develop (#31017)
amp_auto_cast.cc [AMP] Autocast to fp32 for op has no fp16 kernel (#32543)
amp_auto_cast.h [AMP] Autocast to fp32 for op has no fp16 kernel (#32543)
basic_engine.cc clear 'BasicEngine' when an exception occurs in the backward. (#32546)
basic_engine.h add custom init grad for backward function (#31540)
bkcl_context.cc Support control flow in DataParallel (#31625)
bkcl_context.h Support control flow in DataParallel (#31625)
data_loader.cc DataLoader supprot dict str (#31481)
data_loader.h Refine DataLoader support multi-processing (#23107)
dygraph_grad_maker.h Customizable Python Layer in Dygraph (#32130)
engine.h Add dygraph double grad implementation (#22939)
execution_context.h support Exhaustive search in dygraph (#23415)
flags.cc Fix dygraph mem leak (#18082)
flags.h Fix dygraph mem leak (#18082)
gradient_accumulator.cc Add inner register backward hook method for Tensor (#32171)
gradient_accumulator.h Refactor and simplify hook design & add Tensor.register_hook API (#31775)
hooks.h Add inner register backward hook method for Tensor (#32171)
infer_shape_context.h [OpDevOptimize] Add common infershape functions (#26096)
infer_var_type_context.h improve efficiency of runtime InferVarType (#22778)
layer.cc add clearGradient for amp sample code (#32517)
layer.h Customizable Python Layer in Dygraph (#32130)
nccl_context.cc [HybridParallel] fix port reuse when create multi group (#31876)
nccl_context.h [HybridParallel] fix port reuse when create multi group (#31876)
op_base.h Refactor and simplify hook design & add Tensor.register_hook API (#31775)
parallel_context.h Support control flow in DataParallel (#31625)
partial_grad_engine.cc Refactor and simplify hook design & add Tensor.register_hook API (#31775)
partial_grad_engine.h Update the demo code and the doc of varbase.backward. (#26506)
prepared_operator.cc [Custom OP] Support stream set on Custom Op (#31257)
prepared_operator.h add cache for VariableWrapper (#30880)
profiler.cc fix header file paths of gflags, commit 1, test=develop (#30271)
profiler.h 1. Add imperative gperf profiler
py_layer_fwd.h forward return any type. (#32661)
reducer.cc Add inner register backward hook method for Tensor (#32171)
reducer.cu [ROCM] update fluid imperative for rocm (part1), test=develop (#31017)
reducer.h Support control flow in DataParallel (#31625)
saved_variable_wrapper_list.h Add dygraph double grad implementation (#22939)
tracer.cc [Rocm] fix test_var_base (#32639)
tracer.h Customizable Python Layer in Dygraph (#32130)
type_defs.h Add dygraph double grad implementation (#22939)
variable_wrapper.h Customizable Python Layer in Dygraph (#32130)

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

发行版本

当前项目没有发行版本

贡献者 233

全部贡献者

开发语言

  • C++ 47.1 %
  • Python 43.6 %
  • Cuda 7.0 %
  • CMake 1.1 %
  • Shell 0.7 %