You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 
 
 
chenyemeng 8f7de52c12 revert pr 689 3 years ago
..
analyzer !107 bugfix:fix core question because of DestroyGraphJson lack of return 4 years ago
client device-os log missing 3 years ago
common return when task sink failed. 3 years ago
engine_manager ge executor support windows 4 years ago
executor revert pr 689 3 years ago
ge_local_engine 回退 'Pull Request !689 : GeTensor aligned addr & zero copy support' 3 years ago
ge_runtime cpplint magic num & define macro 3 years ago
generator Fix bug of modify output shape to -2. 3 years ago
graph 回退 'Pull Request !689 : GeTensor aligned addr & zero copy support' 3 years ago
host_cpu_engine update dlog change 3 years ago
host_kernels Feature: repair dynamic_stitch_kernel folding bug 3 years ago
hybrid 回退 'Pull Request !689 : GeTensor aligned addr & zero copy support' 3 years ago
inc update code from yellow zone 20201013 4 years ago
init Remove pass reporter callback to hccl 3 years ago
ir_build Add keep_dtype attribute on operators to keep precision unchanged. 3 years ago
model test atc dynamic shape 3 years ago
offline !700 atc --hlep 3 years ago
omm cpplint cast fix 3 years ago
opskernel_manager Remove pass reporter callback to hccl 3 years ago
plugin/engine remove soft link and support other scene cmake 4 years ago
proto add dt_variant type. 3 years ago
session fix cpp lint and errocode 3 years ago
single_op !685 profiling data add iter num 3 years ago
stub profiling AR version 1 3 years ago
CMakeLists.txt revert pr 689 3 years ago
README.md update cmake 4 years ago
ge_inference.mk 回退 'Pull Request !689 : GeTensor aligned addr & zero copy support' 3 years ago
ge_runner.mk 回退 'Pull Request !689 : GeTensor aligned addr & zero copy support' 3 years ago
module.mk update ge folder 4 years ago

图引擎模块(GE)是MindSpore的一个子模块,其代码由C++实现,位于前端模块ME和底层硬件之间,起到承接作用。图引擎模块以ME下发的图作为输入,然后进行一系列的深度图优化操作,最后输出一张可以在底层硬件上高效运行的图。GE针对昇腾AI处理器的硬件结构特点,做了特定的优化工作,以此来充分发挥出昇腾AI处理器的强大算力。在进行模型训练/推理时,GE会被自动调用而用户并不感知。GE主要由GE API和GE Core两部分组成,详细的架构图如下所示

C++ Text Makefile Shell CMake other