提交 · b0d9a3aa526303c4f276e7ef1044ab2913f40aea · 慢慢CG / Mace

28 10月, 2020 1 次提交
- L
  
  fix: Add framework_type arg in Net so that we can get it from Net besides Op · b0d9a3aa
  由 like15 提交于 10月 27, 2020
  
  b0d9a3aa
23 10月, 2020 2 次提交
- L
  feat: add Elu for Tensorflow and Caffe and fix bug in reverse op and net adapter · dd2cdb01
  由 luxuhui 提交于 10月 21, 2020
```
N/A
Signed-off-by: NLuxuhui <luxuhui@xiaomi.com>
```
  dd2cdb01
- L
  
  feat: Support Densenet and MobilenetV2 conversion from PyTorch · 9efe5dc5
  由 like15 提交于 9月 30, 2020
  
  9efe5dc5
22 10月, 2020 1 次提交
- Z
  
  feat: Integrate CMSIS5 to MACE Micro · 6352c291
  由 Zhang Zhimin 提交于 10月 10, 2020
  
  6352c291
27 8月, 2020 1 次提交
- L
  
  Add onnx's Elu operator · c6118a22
  由 lichao18 提交于 8月 27, 2020
  
  c6118a22
22 7月, 2020 1 次提交

Integrate MediaTek APU Support on mt6873, mt6885 and mt6853 (#662) · 63feaf50

由 Yi-Kai-Chen 提交于 7月 22, 2020

* Margaux dev.

* Support Pad, Activation, Mul operator for mnasnet

* Support document enhance model

* Apply add-init-cache-and-preemption patch

* Refine code

* Add APU int16 quantization and fix macro bugs

* Refine code
Co-authored-by: NYungChien Hsu <yungchien.hsu@mediatek.com>
Co-authored-by: NEric Chen <eric-yk.chen@mediatek.com>

63feaf50

17 7月, 2020 2 次提交
- L
  feat: support extract_image_patches op and support dynamic filter of conv2d/deconv2d · 9472dcee
  由 luxuhui 提交于 6月 24, 2020
```
N/A
Signed-off-by: NLuxuhui <luxuhui@xiaomi.com>
```
  9472dcee
- Y
  Integrate MediaTek APU Support on mt6873, mt6885 and mt6853 · 01dbeae3
  由 YungChien Hsu 提交于 3月 11, 2020
```
Signed-off-by: Nlichao18 <lichao18@xiaomi.com>

Add init cache and preemption

Support mnasnet and new 5g library

Add GetModelSize API for model header
```
  01dbeae3
30 6月, 2020 1 次提交
- L
  feat: support splitV op for tensorflow · 041a5d2e
  由 luxuhui 提交于 6月 29, 2020
```
N/A
Signed-off-by: NLuxuhui <luxuhui@xiaomi.com>
```
  041a5d2e
29 6月, 2020 1 次提交
- Y
  
  Add MegEngine converter for MACE (#658) · a28e8128
  由 yzchenmonkey 提交于 6月 29, 2020
  
  a28e8128
17 6月, 2020 1 次提交
- L
  feat: support `half_pixel_centers` arg for `resize` type ops · d24a3ddd
  由 luxuhui 提交于 6月 16, 2020
```
N/A
Signed-off-by: NLuxuhui <luxuhui@xiaomi.com>
```
  d24a3ddd
29 5月, 2020 1 次提交
- L
  feature: support GroupNorm op · 6b65f42b
  由 luxuhui 提交于 5月 12, 2020
```
N/A
Signed-off-by: NLuxuhui <luxuhui@xiaomi.com>
```
  6b65f42b
28 5月, 2020 1 次提交
- L
  support RELU6/ArgMax/ResizeNearestNeighbor op for Caffe, fix bug. · b5dfcf4d
  由 luxuhui 提交于 5月 11, 2020
```
N/A
Signed-off-by: NLuxuhui <luxuhui@xiaomi.com>
```
  b5dfcf4d
09 3月, 2020 1 次提交
- B
  
  Support unet for hexagon · 7c10b286
  由 Bin Li 提交于 3月 04, 2020
  
  7c10b286
07 1月, 2020 1 次提交
- L
  opt the performance of ResizeNearestNeighbor&Deconv OP · c8d5c88e
  由 luxuhui 提交于 1月 06, 2020
```
N/A
Signed-off-by: NLuxuhui <luxuhui@xiaomi.com>
```
  c8d5c88e
06 12月, 2019 1 次提交
- L
  support BatchMatMulV2 & Select ops for tensorflow · 04ecb90e
  由 luxuhui 提交于 12月 04, 2019
```
N/A
Signed-off-by: NLuxuhui <luxuhui@xiaomi.com>
```
  04ecb90e
02 12月, 2019 1 次提交
- L
  
  add tf.sign op and reorder tf ops in tf_converter · c5f92ebf
  由 liutuo 提交于 12月 02, 2019
  
  c5f92ebf
21 11月, 2019 1 次提交
- L
  support kaldi ivector input · d70c7295
  由 liutuo 提交于 7月 09, 2019
```
add input/output for dynamic lstm
use forward indexes for subsample
```
  d70c7295
07 11月, 2019 1 次提交
- L
  add lpnorm、mvnorm op for caffe, enhance biasadd、reshape op · 72a3751f
  由 luxuhui 提交于 10月 25, 2019
```
N/A
Signed-off-by: NLuxuhui <luxuhui@xiaomi.com>
```
  72a3751f
29 10月, 2019 1 次提交

Add CMake and ZH docs · 80a017ce

由 liyin 提交于 9月 27, 2019

Revert "Merge branch 'refactor-python-scripts' into 'master'"

This reverts merge request !1200

80a017ce

28 10月, 2019 4 次提交
- 叶
  
  fix data_format in ReshapeOp · 98924c3a
  由叶剑武提交于 10月 23, 2019
  
  98924c3a
- 叶
  support tf.tile · 021e1e25
  由叶剑武提交于 10月 18, 2019
```
fix type error

fix type warning

add TileOp to MaceTransposableDataFormatOps and reduce memcpy in TileOp
```
  021e1e25
- L
  fix data type convert in onnx-converter · 467aefb0
  由 liutuo 提交于 9月 26, 2019
```
refactor structure of python scripts
```
  467aefb0
- L
  support reshape op on OpenCL · d83cab22
  由 luxuhui 提交于 9月 20, 2019
```
N/A
Signed-off-by: NLuxuhui <luxuhui@xiaomi.com>
```
  d83cab22
10 9月, 2019 1 次提交
- L
  
  Refactor model run: Support CMake everything · 43b33191
  由 liyin 提交于 8月 22, 2019
  
  43b33191
12 8月, 2019 1 次提交
- L
  
  Refactor model converter & encrypt process · 320b509c
  由 liyin 提交于 8月 09, 2019
  
  320b509c
07 8月, 2019 1 次提交
- L
  
  Remove Identity at before and after fake quantize · 02a8c938
  由 liyin 提交于 8月 07, 2019
  
  02a8c938
10 6月, 2019 1 次提交
- L
  
  add reduce sum · 3890a7c5
  由 liutuo 提交于 6月 10, 2019
  
  3890a7c5
31 5月, 2019 1 次提交
- Y
  
  fix gpu squeeze · 7b5a9af2
  由 yejianwu 提交于 5月 31, 2019
  
  7b5a9af2
27 5月, 2019 1 次提交
- L
  
  add unsqueeze constant ops for onnx and fix concat bug · 6d768ae0
  由 liutuo 提交于 5月 27, 2019
  
  6d768ae0
15 5月, 2019 3 次提交
- B
  
  Quantize model for compression · b49c67b8
  由 Bin Li 提交于 5月 09, 2019
  
  b49c67b8
- L
  
  add onnx clip · 405c502a
  由 liutuo 提交于 5月 10, 2019
  
  405c502a
- 許
  Integrate MediaTek APU Support (#440) · 2d650b67
  由許永健提交于 5月 15, 2019
```
* Integrate APU

* fix code style issue
```
  2d650b67
28 4月, 2019 1 次提交
- L
  
  update fp16 · 7b147bbd
  由 liukai6 提交于 4月 12, 2019
  
  7b147bbd
26 4月, 2019 1 次提交
- L
  
  Add data format explanation document and fix bugs in NetDefAdapter. · 95d50602
  由 liuqi 提交于 4月 26, 2019
  
  95d50602
25 4月, 2019 2 次提交
- Y
  
  add FP16_GATHER_WEIGHT · 92f6d362
  由 yulianfei 提交于 11月 29, 2018
  
  92f6d362
- Y
  Reformat fp16 gemv, add FP16_MATMUL_WEIGHT · 0c5aab67
  由 yulianfei 提交于 11月 29, 2018
```
Summary: Reformat fp16 gemv, add FP16_MATMUL_WEIGHT

Differential Revision: https://phabricator.d.xiaomi.net/D144134
```
  0c5aab67
22 4月, 2019 2 次提交
- L
  Update the memory type choose logic and polish some code. · 95b32c24
  由 liuqi 提交于 4月 22, 2019
```
1. Change DataFormat from enum to enum class.
```
  95b32c24
- L
  Refactor: Support auto transformation and net optimization. · 74dcd617
  由 liuqi 提交于 4月 17, 2019
```
1. Auto transformation between data format, data type and memory type.
2. Add device placement optimization strategy.
```
  74dcd617
19 4月, 2019 1 次提交

add kaldi op IfDefined · a7afda0e

由 liutuo 提交于 3月 21, 2019

add extract pooling and kaldi batchnorm

add time index info in extract pooling

add benchmar& test for kaldi components

add batch support for kaldi

change ifdefined to delay

a7afda0e

慢慢CG / Mace 与 Fork 源项目一致

慢慢CG / Mace
与 Fork 源项目一致