Add model card with Apache 2.0 license
Browse files
README.md
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
language:
|
| 4 |
+
- bo
|
| 5 |
+
- zh
|
| 6 |
+
- en
|
| 7 |
+
tags:
|
| 8 |
+
- tibetan-ocr
|
| 9 |
+
- ocr
|
| 10 |
+
- paddleocr
|
| 11 |
+
- tibetan
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
# TibetOCR M5 Model
|
| 15 |
+
|
| 16 |
+
基于 PP-OCRv5 微调的藏文/汉文/英文混排 OCR 识别模型（M5）。
|
| 17 |
+
|
| 18 |
+
## 性能
|
| 19 |
+
|
| 20 |
+
| 测试集 | CA |
|
| 21 |
+
|---|---|
|
| 22 |
+
| 纯藏文 | 99.21% |
|
| 23 |
+
| 藏汉混排 | 97.33% |
|
| 24 |
+
| 藏汉英混排 | 93.52% |
|
| 25 |
+
| 脏污藏文 | 93.65% |
|
| 26 |
+
|
| 27 |
+
## 使用方法
|
| 28 |
+
|
| 29 |
+
详见 [tibetocr 代码仓库](https://github.com/Databaker-tech/tibetocr)。
|
| 30 |
+
|
| 31 |
+
## 许可协议
|
| 32 |
+
|
| 33 |
+
Apache 2.0，可商用、可闭源、可收费，使用时须署名。
|