Upload README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,47 @@
|
|
| 1 |
-
|
| 2 |
-
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Text2Text Legal Clauses Finetuned Model
|
| 2 |
+
|
| 3 |
+
This model fine-tunes google/mt5-small model on shay681/Hugging_Face_df_Legal_Clauses dataset.
|
| 4 |
+
|
| 5 |
+
|
| 6 |
+
## Training and evaluation data
|
| 7 |
+
|
| 8 |
+
| Dataset | Split | # samples |
|
| 9 |
+
| -------- | ----- | --------- |
|
| 10 |
+
| SupremeCourtOfIsrael | train | 147,946 |
|
| 11 |
+
| SupremeCourtOfIsrael | validation | 36,987 |
|
| 12 |
+
|
| 13 |
+
|
| 14 |
+
## Training procedure
|
| 15 |
+
|
| 16 |
+
### Training hyperparameters
|
| 17 |
+
|
| 18 |
+
The following hyperparameters were used during training:
|
| 19 |
+
- evaluation_strategy: "epoch"
|
| 20 |
+
- learning_rate: 5e-5
|
| 21 |
+
- train_batch_size: 4
|
| 22 |
+
- eval_batch_size: 4
|
| 23 |
+
- num_train_epochs: 5
|
| 24 |
+
- weight_decay: 0.01
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
### Framework versions
|
| 28 |
+
|
| 29 |
+
- Transformers 4.17.0
|
| 30 |
+
- Pytorch 1.10.0+cu111
|
| 31 |
+
- Datasets 1.18.4
|
| 32 |
+
- Tokenizers 0.11.6
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
### Results
|
| 36 |
+
|
| 37 |
+
| Metric | # Value |
|
| 38 |
+
| ------ | --------- |
|
| 39 |
+
| **Accuracy** | **0.87** |
|
| 40 |
+
| **F1** | **0.64** |
|
| 41 |
+
|
| 42 |
+
|
| 43 |
+
### About Me
|
| 44 |
+
Created by Shay Doner.
|
| 45 |
+
This is my final project as part of intelligent systems M.Sc studies at Afeka College in Tel-Aviv.
|
| 46 |
+
For more cooperation, please contact email:
|
| 47 |
+
shay681@gmail.com
|