Update README.md
Browse files
README.md
CHANGED
|
@@ -9,8 +9,8 @@ We propose BTC/OTC to directly train an ASR system leveraging weak supervision,
|
|
| 9 |
OTC maintains reasonable ASR performance even when the transcripts contain up to 70% errors of different types.
|
| 10 |
|
| 11 |
|
| 12 |
-
## When transcript error rate = 0.5
|
| 13 |
-
### Results (ctc-greedy-search)
|
| 14 |
<table>
|
| 15 |
<tr>
|
| 16 |
<td rowspan=2>Training Criterion</td>
|
|
@@ -39,7 +39,7 @@ OTC maintains reasonable ASR performance even when the transcripts contain up to
|
|
| 39 |
</tr>
|
| 40 |
</table>
|
| 41 |
|
| 42 |
-
### Results (1best, blank_bias=-4)
|
| 43 |
<table>
|
| 44 |
<tr>
|
| 45 |
<td rowspan=2>Training Criterion</td>
|
|
|
|
| 9 |
OTC maintains reasonable ASR performance even when the transcripts contain up to 70% errors of different types.
|
| 10 |
|
| 11 |
|
| 12 |
+
## When transcript error rate = 0.5
|
| 13 |
+
### Results (WER (%)) (ctc-greedy-search)
|
| 14 |
<table>
|
| 15 |
<tr>
|
| 16 |
<td rowspan=2>Training Criterion</td>
|
|
|
|
| 39 |
</tr>
|
| 40 |
</table>
|
| 41 |
|
| 42 |
+
### Results (WER (%)) (1best, blank_bias=-4)
|
| 43 |
<table>
|
| 44 |
<tr>
|
| 45 |
<td rowspan=2>Training Criterion</td>
|