metadata
license: mit
base_model: microsoft/deberta-v3-base
tags:
- generated_from_trainer
model-index:
- name: legal_deberta
results: []
legal_deberta
This model is a fine-tuned version of microsoft/deberta-v3-base on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.4214
- Law Precision: 0.6449
- Law Recall: 0.92
- Law F1: 0.7582
- Law Number: 75
- Violated by Precision: 0.8625
- Violated by Recall: 0.92
- Violated by F1: 0.8903
- Violated by Number: 75
- Violated on Precision: 0.625
- Violated on Recall: 0.7333
- Violated on F1: 0.6748
- Violated on Number: 75
- Violation Precision: 0.5683
- Violation Recall: 0.6347
- Violation F1: 0.5997
- Violation Number: 616
- Overall Precision: 0.6064
- Overall Recall: 0.6944
- Overall F1: 0.6475
- Overall Accuracy: 0.9475
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 50
Training results
| Training Loss | Epoch | Step | Validation Loss | Law Precision | Law Recall | Law F1 | Law Number | Violated by Precision | Violated by Recall | Violated by F1 | Violated by Number | Violated on Precision | Violated on Recall | Violated on F1 | Violated on Number | Violation Precision | Violation Recall | Violation F1 | Violation Number | Overall Precision | Overall Recall | Overall F1 | Overall Accuracy |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.0391 | 11.11 | 500 | 0.3372 | 0.5652 | 0.8667 | 0.6842 | 75 | 0.8023 | 0.92 | 0.8571 | 75 | 0.6042 | 0.7733 | 0.6784 | 75 | 0.4690 | 0.6640 | 0.5497 | 616 | 0.5141 | 0.7146 | 0.5980 | 0.9283 |
| 0.0036 | 22.22 | 1000 | 0.4019 | 0.5667 | 0.9067 | 0.6974 | 75 | 0.7955 | 0.9333 | 0.8589 | 75 | 0.5455 | 0.72 | 0.6207 | 75 | 0.5681 | 0.6429 | 0.6032 | 616 | 0.5857 | 0.6992 | 0.6374 | 0.9443 |
| 0.0002 | 33.33 | 1500 | 0.3958 | 0.6 | 0.92 | 0.7263 | 75 | 0.8023 | 0.92 | 0.8571 | 75 | 0.5556 | 0.7333 | 0.6322 | 75 | 0.5476 | 0.6347 | 0.5880 | 616 | 0.5759 | 0.6944 | 0.6296 | 0.9463 |
| 0.0001 | 44.44 | 2000 | 0.4214 | 0.6449 | 0.92 | 0.7582 | 75 | 0.8625 | 0.92 | 0.8903 | 75 | 0.625 | 0.7333 | 0.6748 | 75 | 0.5683 | 0.6347 | 0.5997 | 616 | 0.6064 | 0.6944 | 0.6475 | 0.9475 |
Framework versions
- Transformers 4.31.0
- Pytorch 2.4.0+cu121
- Datasets 2.15.0
- Tokenizers 0.13.3