Update README.md
Browse files
README.md
CHANGED
|
@@ -43,7 +43,13 @@ This model is a fine-tuned version of [google/mt5-large](https://huggingface.co/
|
|
| 43 |
|
| 44 |
#### Training Hyperparameters
|
| 45 |
|
| 46 |
-
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 47 |
|
| 48 |
#### Metrics
|
| 49 |
|
|
|
|
| 43 |
|
| 44 |
#### Training Hyperparameters
|
| 45 |
|
| 46 |
+
- learning_rate: 5e-05
|
| 47 |
+
- train_batch_size: 1
|
| 48 |
+
- gradient_accumulation_steps: 16
|
| 49 |
+
- total_train_batch_size: 16
|
| 50 |
+
- optimizer: AdamW
|
| 51 |
+
- lr_scheduler_type: linear
|
| 52 |
+
- max_grad_norm: 0.3
|
| 53 |
|
| 54 |
#### Metrics
|
| 55 |
|