Skip to content

Commit 7d9ff40

Browse files
committed
narrow pictures
1 parent 02e0b5f commit 7d9ff40

File tree

6 files changed

+10
-10
lines changed

6 files changed

+10
-10
lines changed

doc/design/mkldnn/README.MD

Lines changed: 10 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -5,7 +5,7 @@
55
充分展现英特尔平台的优势,有效提升PaddlePaddle在英特尔架构上的性能。
66

77
<div align="center">
8-
<img src="image/overview.png" width=410><br/>
8+
<img src="image/overview.png"><br/>
99
Figure 1. PaddlePaddle on IA
1010
</div>
1111

@@ -42,16 +42,16 @@ Figure 1. PaddlePaddle on IA
4242

4343
MKL,MKLML以及MKL-DNN三者关系如下表:
4444

45-
| Name | Open Source | License | Descriptions |
46-
|------------|----------------| ------------| --------------|
47-
| MKL | No | Proprietary | Accelerate math processing routines |
48-
| MKLML | No | Proprietary | Small package of MKL, especially for Machine Learning |
49-
| MKL-DNN | Yes | Apache 2.0 | Accelerate primitives processing routines especially for Deep Neural Networks |
45+
| Name | Open Source | License | Descriptions |
46+
| :---------- | :--------------- | :---------- | :------------ |
47+
| MKL | No | Proprietary | Accelerate math processing routines |
48+
| MKLML | No | Proprietary | Small package of MKL, especially for Machine Learning |
49+
| MKL-DNN | Yes | Apache 2.0 | Accelerate primitives processing routines especially for Deep Neural Networks |
5050

5151
MKLML可以与MKL-DNN共同使用,以此达到最好的性能。
5252

5353
<div align="center">
54-
<img src="image/engine.png" width=410><br/>
54+
<img src="image/engine.png"><br/>
5555
Figure 2. PaddlePaddle with MKL Engines
5656
</div>
5757

@@ -103,7 +103,7 @@ MKL-DNN的库目前只有动态库`libmkldnn.so`。
103103
所以我们定义了一个`MKLDNNMatrix`用于管理MKL-DNN数据的不同格式以及相互之间的转换。
104104

105105
<div align="center">
106-
<img src="image/matrix.png" height=300><br/>
106+
<img src="image/matrix.png"><br/>
107107
Figure 3. MKLDNNMatrix
108108
</div>
109109

@@ -113,7 +113,7 @@ Figure 3. MKLDNNMatrix
113113
子类只需要使用定义好的接口,实现具体的函数功能即可。
114114

115115
<div align="center">
116-
<img src="image/layers.png" width=410><br/>
116+
<img src="image/layers.png"><br/>
117117
Figure 4. MKLDNNLayer
118118
</div>
119119

@@ -150,7 +150,7 @@ Figure 4. MKLDNNLayer
150150
所以整体上,在实现每个子类的时候就不需要关心分支的事情了。
151151

152152
<div align="center">
153-
<img src="image/gradients.png" height=300><br/>
153+
<img src="image/gradients.png"><br/>
154154
Figure 5. Merge Gradients
155155
</div>
156156

doc/design/mkldnn/image/engine.png

-3.43 KB
Loading

doc/design/mkldnn/image/gradients.png

-8.16 KB
Loading

doc/design/mkldnn/image/layers.png

-2.7 KB
Loading

doc/design/mkldnn/image/matrix.png

-3.59 KB
Loading

doc/design/mkldnn/image/overview.png

-5.43 KB
Loading

0 commit comments

Comments
 (0)