Add WER with LM on train data
Browse files
README.md
CHANGED
|
@@ -48,6 +48,21 @@ It achieves the following results on the evaluation set:
|
|
| 48 |
- Loss: 0.3119
|
| 49 |
- Wer: 0.2613
|
| 50 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 51 |
## Model description
|
| 52 |
|
| 53 |
More information needed
|
|
@@ -114,19 +129,6 @@ The following hyperparameters were used during training:
|
|
| 114 |
| 0.3133 | 149.04 | 15500 | 0.3114 | 0.2624 |
|
| 115 |
|
| 116 |
|
| 117 |
-
### Evaluation metrics
|
| 118 |
-
|
| 119 |
-
| Metric | Split | LM | Value |
|
| 120 |
-
|:------:|:------:|:-----:|:---------:|
|
| 121 |
-
| WER | Train | No | 5.36 |
|
| 122 |
-
| CER | Train | No | 1.11 |
|
| 123 |
-
| WER | Test | No | 26.14 |
|
| 124 |
-
| CER | Test | No | 4.93 |
|
| 125 |
-
| WER | Train | Yes | |
|
| 126 |
-
| CER | Train | Yes | |
|
| 127 |
-
| WER | Test | Yes | 20.69 |
|
| 128 |
-
| CER | Test | Yes | 3.986 |
|
| 129 |
-
|
| 130 |
### Framework versions
|
| 131 |
|
| 132 |
- Transformers 4.16.0.dev0
|
|
|
|
| 48 |
- Loss: 0.3119
|
| 49 |
- Wer: 0.2613
|
| 50 |
|
| 51 |
+
|
| 52 |
+
### Evaluation metrics
|
| 53 |
+
|
| 54 |
+
| Metric | Split | Decode with LM | Value |
|
| 55 |
+
|:------:|:------:|:--------------:|:---------:|
|
| 56 |
+
| WER | Train | No | 5.36 |
|
| 57 |
+
| CER | Train | No | 1.11 |
|
| 58 |
+
| WER | Test | No | 26.14 |
|
| 59 |
+
| CER | Test | No | 4.93 |
|
| 60 |
+
| WER | Train | Yes | 5.04 |
|
| 61 |
+
| CER | Train | Yes | 1.07 |
|
| 62 |
+
| WER | Test | Yes | 20.69 |
|
| 63 |
+
| CER | Test | Yes | 3.986 |
|
| 64 |
+
|
| 65 |
+
|
| 66 |
## Model description
|
| 67 |
|
| 68 |
More information needed
|
|
|
|
| 129 |
| 0.3133 | 149.04 | 15500 | 0.3114 | 0.2624 |
|
| 130 |
|
| 131 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 132 |
### Framework versions
|
| 133 |
|
| 134 |
- Transformers 4.16.0.dev0
|