Commit
•
db8da66
1
Parent(s):
00d2d81
Update README.md
Browse files
README.md
CHANGED
|
@@ -39,7 +39,7 @@ It achieves the following results on the evaluation set:
|
|
| 39 |
## Training procedure
|
| 40 |
It was first trained 2000 steps on Lithuanian and then 2000 steps on Serbian, continuing from the last checkpoint for Lithuanian.
|
| 41 |
|
| 42 |
-
### Training hyperparameters
|
| 43 |
|
| 44 |
The following hyperparameters were used during training:
|
| 45 |
- learning_rate: 1e-05
|
|
@@ -52,14 +52,6 @@ The following hyperparameters were used during training:
|
|
| 52 |
- training_steps: 2000
|
| 53 |
- mixed_precision_training: Native AMP
|
| 54 |
|
| 55 |
-
### Training results for Serbian
|
| 56 |
-
|
| 57 |
-
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 58 |
-
|:-------------:|:-----:|:----:|:---------------:|:-------:|
|
| 59 |
-
| 0.0003 | 49.01 | 1000 | 0.6035 | 26.7908 |
|
| 60 |
-
| 0.0002 | 99.01 | 2000 | 0.6311 | 35.6131 |
|
| 61 |
-
|
| 62 |
-
|
| 63 |
### Framework versions
|
| 64 |
|
| 65 |
- Transformers 4.26.0.dev0
|
|
|
|
| 39 |
## Training procedure
|
| 40 |
It was first trained 2000 steps on Lithuanian and then 2000 steps on Serbian, continuing from the last checkpoint for Lithuanian.
|
| 41 |
|
| 42 |
+
### Training hyperparameters per fine-tune
|
| 43 |
|
| 44 |
The following hyperparameters were used during training:
|
| 45 |
- learning_rate: 1e-05
|
|
|
|
| 52 |
- training_steps: 2000
|
| 53 |
- mixed_precision_training: Native AMP
|
| 54 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 55 |
### Framework versions
|
| 56 |
|
| 57 |
- Transformers 4.26.0.dev0
|