Update README.md
Browse files
README.md
CHANGED
|
@@ -1,7 +1,6 @@
|
|
| 1 |
---
|
| 2 |
language:
|
| 3 |
- en
|
| 4 |
-
license: apache-2.0
|
| 5 |
datasets:
|
| 6 |
- Open-Orca/OpenOrca
|
| 7 |
- jackhhao/jailbreak-classification
|
|
@@ -30,9 +29,16 @@ Fine-tuned on the [jailbreak-classification](https://huggingface.co/datasets/jac
|
|
| 30 |
|
| 31 |
#### Training Hyperparameters
|
| 32 |
|
| 33 |
-
Fine-tuning hyper-parameters
|
| 34 |
- learning_rate = 5e-5
|
| 35 |
- train_batch_size = 8
|
| 36 |
- eval_batch_size = 8
|
| 37 |
- lr_scheduler_type = linear
|
| 38 |
-
- num_train_epochs = 5.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
language:
|
| 3 |
- en
|
|
|
|
| 4 |
datasets:
|
| 5 |
- Open-Orca/OpenOrca
|
| 6 |
- jackhhao/jailbreak-classification
|
|
|
|
| 29 |
|
| 30 |
#### Training Hyperparameters
|
| 31 |
|
| 32 |
+
Fine-tuning hyper-parameters(on train(0.8) and val(0.2))
|
| 33 |
- learning_rate = 5e-5
|
| 34 |
- train_batch_size = 8
|
| 35 |
- eval_batch_size = 8
|
| 36 |
- lr_scheduler_type = linear
|
| 37 |
+
- num_train_epochs = 5.0
|
| 38 |
+
|
| 39 |
+
second fine-tuning hyper-parameters(on train and test)
|
| 40 |
+
- learning_rate = 1e-5
|
| 41 |
+
- train_batch_size = 8
|
| 42 |
+
- eval_batch_size = 8
|
| 43 |
+
- lr_scheduler_type = linear
|
| 44 |
+
- num_train_epochs = 3.0
|