p09823145
/
InfiniTensor
forked from jiuyuan/InfiniTensor
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
238 Commits 70 Branches 0 Tags 141 MiB
Commit Graph

238 Commits

Author SHA1 Message Date
Hardy 823e66a9ff
Support perf bang 1115 (#57) 
* support matmul

* add matmul

* add matmul

* add code for cnnl matmul operation and test

* add conv

* add code for conv test on mlu

* add code for test cnnl conv on mlu

* add code for perf conv and matmul on mlu

* clang format

* fix convolution operation

* fxi cmaklist

* code format

* fix code

* code format

---------

Co-authored-by: wanghailu <wanghailu@qiyuanlab.com>
Co-authored-by: wanghailu <wanghailu0717@163.com>
 2023-03-29 13:52:56 +08:00
wendy12022 86ec4036ce
ADD: add mkl runtime for intel cpu , and add mkl kernel for matmul/conv/convtransposed. (#61) 
* move memory format transformation to TensorObj

clang format

add MemoryFormat for tensorObj.

use post_ops for fused conv/deconv

Distinguish mkl  op_timer from cuda op timer.

add act optype to conv and deconv

add operator timer

add mkl kernel for convTransposed

minor fix for group conv

do not use cblas_sgemm_batch

CpuRuntimeObj->NativeCpuRuntimeObj

add  matmul op for mkl

* fix: fix bugs when rebasing from master

fix: fix bugs when rebasing from master

* fix: update api after rebasing

* fix: fix format; fix onnx import

* fix: fix clang-format

* [fix] fix conv_transpose test

* [fix] use stronger test case for transposed conv

* [fix] remove tensor memory format; fix mkl transpose conv

* [fix] add FIXME tag for op_timer python api

---------

Co-authored-by: whjthu <haojie0429@gmail.com>
 2023-03-27 21:28:49 +08:00
Haojie Wang 65a3abf5dc
feat: inference (#71) 
导出推理接口，支持通过 python 调用框架推理
 2023-03-25 12:09:22 +08:00
whjthu d9886e9de3 fix: remove inline keyword in class; rename getter and setter for inputOf and outputOf 2023-03-25 12:04:24 +08:00
YdrMaster aff2b538ce fix: 删除单独的拷贝函数 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-22 10:13:06 +08:00
wanghailu 64a5de51f3 fix 2023-03-22 10:08:31 +08:00
YdrMaster 5aeacedab3 fix: 从模板导出每个类型的 python 接口 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-22 09:46:40 +08:00
YdrMaster 73e895b8ce feat: 导出拷出张量值方法 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-21 14:40:13 +08:00
YdrMaster 9db97eb212 refactor: 整合操作张量数据的方法 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-21 14:00:04 +08:00
YdrMaster e1c976568d fix: 增加推理接口 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster c18845a2fd feat: 增加推理接口 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster 6e1af09dd0 fix: remove print 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster e294e46436 feat: 导出 pool 到 onnx 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster 8a871c3773 feat: 导出 conv 到 onnx 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster afed749b74 feat: 支持导出权重 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster 40fb8390b1 feat: 导入时保存权重 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster a5e692baea feat: 导出 batchnorm 到 onnx 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster 71ca4459d9 fmt 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster 5b6698bac7 feat: 导出全图的输出张量到 onnx 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster 59bf59c10b docs: update README.md 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster fb3478bf3e build: update Makefile 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
wanghailu 514666591e add batch_norm 2023-03-15 17:23:32 +08:00
YdrMaster 3d122aebfe feat: 支持导出浮点向量 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster cf9bdb0562 feat: 支持打印结果 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster f44a4daf70 feat: 导出未初始化的张量 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster 6dce129cb3 fix: TensorObj::dataMalloc 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster dc79b72655 fix: 重新导出 cuda_runtime() 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster 9ab78f13f7 feat: 导出 cuda_runtime 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster 60c5d6b5b8 fix: 先不在 cpu 上测试 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster ed81861375 temp: 实现初始值导入，但 resnet 报错 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster 4ffaa44c1e fix: Matmul 支持 2 维或以上的输入 
> 现在能导入 resnet18

Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
YdrMaster a27391fcdc fix: 修正 batchNorm 实现 
- onnx 和 pytorch 认为 batchNorm 的 4 个参数是 [c] 形状的，cuDNN 可能认为是 [1,c,1,...]。
优化已改为 [c]，但 cuDNN 推理没有改；

Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 17:23:32 +08:00
Haojie Wang dd5d091dbc
wip: onnx 导出 (#65) 
| Notice | Work in progress
|-|-

> based on #63 

## Progress

1. [x] 对节点拓扑排序
2. [x] 遍历节点，命名并导出其输出张量（`make_tensor_value_info`）
3. [x] 识别图的输入张量，命名并导出（`make_tensor_value_info`）
4. [x] 根据节点类型，识别权重及属性，导出节点（`make_node`）
5. [x] `make_graph` -> `check_graph` -> `make_model` -> `check_model`
 2023-03-15 15:22:09 +08:00
YdrMaster 62fd619987 fix: 移除中文注释 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 15:16:16 +08:00
YdrMaster 71a87c27d1 feat: 导出 ReduceMean 到 onnx 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 15:09:12 +08:00
YdrMaster bb9b62b169 fix: 改正类型转换 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 15:09:12 +08:00
YdrMaster 2a23669394 feat: 导出 Reshape 到 onnx 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 15:09:12 +08:00
YdrMaster ffd0473bd2 feat: check everything 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 15:09:12 +08:00
YdrMaster 9e0f8f21bf feat: 生成模型对象 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 15:09:12 +08:00
YdrMaster 6b7af7077b feat: 导出 Gather Concat 到 onnx 
- 并优化 python 代码

Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 15:09:12 +08:00
YdrMaster 9d9fbd44af feat: 导出 MatMul Concat 到 onnx 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 15:09:12 +08:00
YdrMaster 32f6f02c81 feat: 导出 5 个单目算子到 onnx 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 15:09:12 +08:00
YdrMaster 0517089dca feat: 导出输入张量到 onnx 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 15:09:12 +08:00
YdrMaster eff4c14a85 feat: 封装上下文对象以复用建图代码 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 15:09:12 +08:00
YdrMaster 0833a2f779 feat: 导出加减乘除幂到 onnx 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 15:09:12 +08:00
YdrMaster f2591edbb4 feat: 导出 OpType，为节点命名 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 15:09:12 +08:00
YdrMaster fe81fccf76 feat: 导出 OperatorObj 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 15:09:12 +08:00
YdrMaster 45a3cdfa30 feat: GraphObj 增加一个拓扑排序方法及其测试 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 15:09:12 +08:00
YdrMaster f20e791cf5 style: 修改 graph.h/graph.cc 
Signed-off-by: YdrMaster <ydrml@hotmail.com>
 2023-03-15 15:09:12 +08:00
Haojie Wang 0f52d04882
Merge branch 'master' into dev-onnx 2023-03-15 14:52:03 +08:00