提交 · a83e847fab3f9479fd681ab8d1fbbfdb8876c633 · PaddlePaddle / Paddle-Lite

15 3月, 2020 4 次提交

[LITE][OPENCL] Add lws for opencl conv image kernel (#3191) · a83e847f

由 Yuan Shuai 提交于 3月 15, 2020

* [LITE][OPENCL] Change fp32 fc to fp16's. test=develop

* fix act in conv3x3opt opencl kernel. test=develop

* [LITE][OPENCL] fix opencl fc kernel. test=develop

* [LITE][OPENCL] add lws for opencl conv image kernel. test=develop

a83e847f

[LITE][OPENCL] Fix fc opencl (#3190) · 8fe9a566

由 Yuan Shuai 提交于 3月 15, 2020

* [LITE][OPENCL] Change fp32 fc to fp16's. test=develop

* fix act in conv3x3opt opencl kernel. test=develop

* [LITE][OPENCL] fix opencl fc kernel. test=develop

8fe9a566

X

[LITE][OPENCL][Image] open depthwise spl ,test=develop (#3186) · 6fec9665
由 xiebaiyuan 提交于 3月 15, 2020

6fec9665

[LITE][OPENCL] Fix armv7 gcc opencl compile (#3185) · 6617aa8b

由 Yuan Shuai 提交于 3月 15, 2020

* [LITE][OPENCL] Change fp32 fc to fp16's. test=develop

* fix act in conv3x3opt opencl kernel. test=develop

* [LITE][OPENCL] Fix issue(previous definition in gcc.a) armv7 gcc compile for opencl. test=develop

6617aa8b

14 3月, 2020 7 次提交
- H
  [arm] improve conv3x3_dw performance with relu, relu6 and leakey relu (#3183) · 14e43112
  由 HappyAngel 提交于 3月 14, 2020
```
* improve conv_dw profile with rel relu6 leakyrelu, test=develop

* add depthwise, test=develop

* fix ci error, test=develop

* fix cv demo print, test=develop
```
  14e43112
- Y
  [ARM] improve sgemm performance with relu, relu6 and leakey relu, test=develop (#3181) · f7f65134
  由 yiicy 提交于 3月 14, 2020
```
* [ARM] improve sgemm performance with relu, relu6 and leakey relu, test=develop

* [ARM] improve sgemm performance with relu, relu6 and leakey relu, test=develop
```
  f7f65134
- H
  [cv] fix resize error (#3178) · 50822e44
  由 HappyAngel 提交于 3月 14, 2020
```
* fix resie bug

* fix format, test=develop

* fix memory bug, test=develop

* fix format. test=develop
```
  50822e44
- Y
  [LITE][DEMO] Fix mobile light demo (#3180) · bb217292
  由 Yuan Shuai 提交于 3月 14, 2020
```
* [LITE][OPENCL] Change fp32 fc to fp16's. test=develop

* fix act in conv3x3opt opencl kernel. test=develop

* [LITE][DEMO] fix input arg. test=develop

* fix shutdownlog for std::endl. test=develop
```
  bb217292
- J
  
  [Lite][OpenCL]Delete unused macro in depthwise_conv2d_kernel.cl. test=develop (#3176) · fcdfb596
  由 Jiaying Zhao 提交于 3月 14, 2020
  
  fcdfb596
- Y
  [LITE][OPENCL] Change fp32 fc to fp16's. test=develop (#3173) · 475187cc
  由 Yuan Shuai 提交于 3月 14, 2020
```
* [LITE][OPENCL] Change fp32 fc to fp16's. test=develop

* fix act in conv3x3opt opencl kernel. test=develop
```
  475187cc
- Y
  [LITE][DEMO] Enhance mobile_light demo. test=develop (#3171) · 27e90303
  由 Yuan Shuai 提交于 3月 14, 2020
```
* [LITE][DEMO] Enhance mobile_light demo. test=develop

* fix print. test=develop
```
  27e90303
13 3月, 2020 4 次提交
- J
  
  [Lite][OpenCL]Add conv2d_3x3_opt_kernel. test=develop (#3170) · a8e62ae8
  由 Jiaying Zhao 提交于 3月 13, 2020
  
  a8e62ae8
- X
  
  fix: remove winograd extend_workspace to run (#3168) · e9989afd
  由 xiaogang 提交于 3月 13, 2020
  
  e9989afd
- H
  【arm】resize nv12 bug (#3155) · ba54e537
  由 HappyAngel 提交于 3月 13, 2020
```
* fix clang v7 error
* change ut, test=develop
* fix conv ut, test=develop
```
  ba54e537
- H
  
  [opt]add precision type into opt test=develop (#3151) · 3501f3c2
  由 huzhiqiang 提交于 3月 13, 2020
  
  3501f3c2
12 3月, 2020 3 次提交
- H
  [Core] Fix findar for Clang (#3160) · c7fb4138
  由 hong19860320 提交于 3月 12, 2020
```
* [Core] Fix findar for Clang
test=develop

* [CORE] Fix the missing of feed and fetch kernel when using light API to loade the optimized model
test=develop
```
  c7fb4138
- M
  add mean op (#3130) · b9273631
  由 mapingshuo 提交于 3月 12, 2020
```
* add mean op, test=develop

* split forward and backward, test=develop
```
  b9273631
- Y
  
  fix clang compile opencl. test=develop (#3153) · 054fd1af
  由 Yuan Shuai 提交于 3月 12, 2020
  
  054fd1af
11 3月, 2020 2 次提交
- Y
  
  [MOBILE][OPENCL] Enhance mobilenet gpu demo. test=develop (#3156) · fb929405
  由 Yuan Shuai 提交于 3月 11, 2020
  
  fb929405
- Y
  [OPENCL] add pad2d image kernel and ut, test=develop (#3143) · 0f295c87
  由 yiicy 提交于 3月 11, 2020
```
add pad2d image kernel and ut
```
  0f295c87
10 3月, 2020 11 次提交
- H
  
  [CORE] Support the fully quantized model for MTK and RK NPU (#3096) · 08a3ed12
  由 hong19860320 提交于 3月 10, 2020
  
  08a3ed12
- X
  【opencl】add dropout opencl kernel (#3141) · 89da9953
  由 xiaogang 提交于 3月 10, 2020
```
* feat: add dropout opencl kernel
```
  89da9953
- C
  
  Add library and opt requirement for post_quant_without_dat, test=develop, test=document_fix (#3140) · b3653b7e
  由 cc 提交于 3月 10, 2020
  
  b3653b7e
- H
  [ARM]fix conv3x3_dw v7 clang bug (#3129) · a38dc43b
  由 HappyAngel 提交于 3月 10, 2020
```
* fix v7 clang bug. test=develop

* fix format, test=develop

* fix .a build bug, test=develop
```
  a38dc43b
- C
  
  Add import for post_quant_without_data doc, test=develop, test=document_fix (#3137) · cbeff2b5
  由 cc 提交于 3月 10, 2020
  
  cbeff2b5
- C
  
  Adjust the framework of doc, test=develop, test=document_fix (#3134) · 3bc6a928
  由 cc 提交于 3月 10, 2020
  
  3bc6a928
- Z
  [NPU] support more elementwise cases (#3133) · a981af15
  由 zhupengyang 提交于 3月 10, 2020
```
* [NPU] revert shape check for input&output

* [NPU] reshape elt input&output
```
  a981af15
- X
  [LITE][OPENCL] add slice kernel (#3126) · 66f0b25b
  由 xiaogang 提交于 3月 10, 2020
```
* feat: add opencl elementwise_sub op & ut
```
  66f0b25b
- 石
  
  update lite.map (#3127) · 34c29406
  由石晓伟提交于 3月 10, 2020
  
  34c29406
- Y
  [LITE][BUILD] fix clang compile armv8 cpu 's shared lib. test=develop (#3123) · 5ec1a087
  由 Yuan Shuai 提交于 3月 10, 2020
```
* fix clang compile armv8 cpu 's shared lib. test=develop

* use build-in shared helper function. test=develop
```
  5ec1a087
- X
  
  [mobile][opencl] fix conv add bn relu ,test=develop (#3131) · 4e354279
  由 xiebaiyuan 提交于 3月 10, 2020
  
  4e354279
09 3月, 2020 9 次提交
- Y
  [OPENCL] add instance norm kernel and ut, test=develop (#3122) · 604c5d4f
  由 yiicy 提交于 3月 09, 2020
```
add instance norm kernel and ut
```
  604c5d4f
- H
  [arm] fix clang v7 bug (#3118) · 1bd34c3f
  由 HappyAngel 提交于 3月 09, 2020
```
* set arm_lang default is off. test=develop

* fix resize error, test-develop
```
  1bd34c3f
- M
  add lookup_dequant_op (#3108) · f427eb18
  由 mapingshuo 提交于 3月 09, 2020
```
* add lookup_dequant_op
```
  f427eb18
- Z
  
  avoid reusing non-tensor in memery_optimize_pass (#3111) · e0bf1152
  由 zhupengyang 提交于 3月 09, 2020
  
  e0bf1152
- H
  
  [Android demo docs]modify android demo doc (#3119) · 736de982
  由 huzhiqiang 提交于 3月 09, 2020
  
  736de982
- M
  
  fix sgemm bug, test=develop (#3053) · 43a371fe
  由 mapingshuo 提交于 3月 09, 2020
  
  43a371fe
- X
  Opencl eltwisesub (#3116) · 706a44d8
  由 xiaogang 提交于 3月 09, 2020
```
* feat: add opencl elementwise_sub op & ut
```
  706a44d8
- Z
  enhance kernels of transformer decoder (#3110) · b086e835
  由 zhupengyang 提交于 3月 09, 2020
```
* enhance gather, lookup_table arm kernel uts

* enhance beam_search, beam_search_decoder, increment rigster
```
  b086e835
- X
  
  [LITE][OPENCL][Image] opencl compatible with arm ,test=develop (#3120) · e2a02f63
  由 xiebaiyuan 提交于 3月 09, 2020
  
  e2a02f63