Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
Menu
Open sidebar
wangsen
paddle_dbnet
Commits
e37c6b2b
Unverified
Commit
e37c6b2b
authored
May 07, 2022
by
xiaoting
Committed by
GitHub
May 07, 2022
Browse files
update arabic doc and ppocr_v3 doc (#6162)
parent
16173562
Changes
3
Show whitespace changes
Inline
Side-by-side
Showing
3 changed files
with
13 additions
and
11 deletions
+13
-11
doc/doc_ch/PP-OCRv3_introduction.md
doc/doc_ch/PP-OCRv3_introduction.md
+13
-11
doc/imgs_words/arabic/ar_1.jpg
doc/imgs_words/arabic/ar_1.jpg
+0
-0
doc/imgs_words/arabic/ar_2.jpg
doc/imgs_words/arabic/ar_2.jpg
+0
-0
No files found.
doc/doc_ch/PP-OCRv3_introduction.md
View file @
e37c6b2b
...
@@ -110,14 +110,15 @@ PP-OCRv3识别模型从网络结构、训练策略、数据增广等多个方面
...
@@ -110,14 +110,15 @@ PP-OCRv3识别模型从网络结构、训练策略、数据增广等多个方面
|-----|-----|--------|----| --- |
|-----|-----|--------|----| --- |
| 01 | PP-OCRv2 | 8M | 74.8% | 8.54ms |
| 01 | PP-OCRv2 | 8M | 74.8% | 8.54ms |
| 02 | SVTR_Tiny | 21M | 80.1% | 97ms |
| 02 | SVTR_Tiny | 21M | 80.1% | 97ms |
| 03 | SVTR_LCNet | 12M | 71.9% | 6.6ms |
| 03 | SVTR_LCNet(h32) | 12M | 71.9% | 6.6ms |
| 04 | + GTC | 12M | 75.8% | 7.6ms |
| 04 | SVTR_LCNet(h48) | 12M | 73.98% | 7.6ms |
| 05 | + TextConAug | 12M | 76.3% | 7.6ms |
| 05 | + GTC | 12M | 75.8% | 7.6ms |
| 06 | + TextRotNet | 12M | 76.9% | 7.6ms |
| 06 | + TextConAug | 12M | 76.3% | 7.6ms |
| 07 | + UDML | 12M | 78.4% | 7.6ms |
| 07 | + TextRotNet | 12M | 76.9% | 7.6ms |
| 08 | + UIM | 12M | 79.4% | 7.6ms |
| 08 | + UDML | 12M | 78.4% | 7.6ms |
| 09 | + UIM | 12M | 79.4% | 7.6ms |
注: 测试速度时,实验01-03输入图片尺寸均为(3,32,320),04-0
8
输入图片尺寸均为(3,48,320)。在实际预测时,图像为变长输入,速度会有所变化。
注: 测试速度时,实验01-03输入图片尺寸均为(3,32,320),04-0
9
输入图片尺寸均为(3,48,320)。在实际预测时,图像为变长输入,速度会有所变化。
**(1)轻量级文本识别网络SVTR_LCNet。**
**(1)轻量级文本识别网络SVTR_LCNet。**
...
@@ -153,9 +154,10 @@ PP-OCRv3将base模型从CRNN替换成了[SVTR](https://arxiv.org/abs/2205.00159)
...
@@ -153,9 +154,10 @@ PP-OCRv3将base模型从CRNN替换成了[SVTR](https://arxiv.org/abs/2205.00159)
| 02 | SVTR_Tiny | 21M | 80.1% | 97ms |
| 02 | SVTR_Tiny | 21M | 80.1% | 97ms |
| 03 | SVTR_LCNet(G4) | 9.2M | 76% | 30ms |
| 03 | SVTR_LCNet(G4) | 9.2M | 76% | 30ms |
| 04 | SVTR_LCNet(G2) | 13M | 72.98% | 9.37ms |
| 04 | SVTR_LCNet(G2) | 13M | 72.98% | 9.37ms |
| 05 | SVTR_LCNet | 12M | 71.9% | 6.6ms |
| 05 | SVTR_LCNet(h32) | 12M | 71.9% | 6.6ms |
| 06 | SVTR_LCNet(h48) | 12M | 73.98% | 7.6ms |
注: 测试速度时,输入图片尺寸均为(3,32,320); PP-OCRv2-baseline 代表没有借助蒸馏方法训练得到的模型
注: 测试速度时,
01-05
输入图片尺寸均为(3,32,320); PP-OCRv2-baseline 代表没有借助蒸馏方法训练得到的模型
**(2)采用Attention指导CTC训练。**
**(2)采用Attention指导CTC训练。**
...
...
doc/imgs_words/arabic/ar_1.jpg
View replaced file @
16173562
View file @
e37c6b2b
4.7 KB
|
W:
|
H:
46.9 KB
|
W:
|
H:
2-up
Swipe
Onion skin
doc/imgs_words/arabic/ar_2.jpg
View replaced file @
16173562
View file @
e37c6b2b
3.62 KB
|
W:
|
H:
13.1 KB
|
W:
|
H:
2-up
Swipe
Onion skin
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment