Commit
·
89d5a75
1
Parent(s):
c15601e
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,21 @@
|
|
| 1 |
---
|
| 2 |
license: mit
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: mit
|
| 3 |
+
language:
|
| 4 |
+
- zh
|
| 5 |
+
pipeline_tag: image-to-text
|
| 6 |
---
|
| 7 |
+
|
| 8 |
+
Convert images of IPA phonetic symbols to Pinyin.
|
| 9 |
+
|
| 10 |
+
# Target: Convert Scanned IPA symbols to Pinyin
|
| 11 |
+
Scanned images of IPA phonetic symbols for Chengdunese (成都话) in The Great Dictionary of Modern Chinese Dialects (現代漢語方言大詞典).
|
| 12 |
+
|
| 13 |
+
TODO: labeled part of the test set.
|
| 14 |
+
|
| 15 |
+
# Training and Test Set
|
| 16 |
+
* 2,553 images of IPA phonetic symbols generated from Pinyin pronunciations found in Sichuanese Dialect Dictionary (四川方言词典 教你一口地道的四川话) and the word list of the Shupin (蜀拼) input method.
|
| 17 |
+
* 80/20 split on train/test
|
| 18 |
+
|
| 19 |
+
# Results
|
| 20 |
+
* Trained for 180 steps with a batch size of 32
|
| 21 |
+
* Final Character Error Rate of 0.795%
|