File size: 6,591 Bytes
64f60c4 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 |
---
base_model: Qwen/Qwen2-7B
library_name: peft
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: legal_qwen
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# legal_qwen
This model is a fine-tuned version of [Qwen/Qwen2-7B](https://huggingface.co/Qwen/Qwen2-7B) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5566
- Law Precision: 0.4434
- Law Recall: 0.6351
- Law F1: 0.5222
- Law Number: 74
- Violated by Precision: 0.3663
- Violated by Recall: 0.5139
- Violated by F1: 0.4277
- Violated by Number: 72
- Violated on Precision: 0.1507
- Violated on Recall: 0.2245
- Violated on F1: 0.1803
- Violated on Number: 49
- Violation Precision: 0.3132
- Violation Recall: 0.4430
- Violation F1: 0.3669
- Violation Number: 596
- Overall Precision: 0.3197
- Overall Recall: 0.4539
- Overall F1: 0.3751
- Overall Accuracy: 0.9062
## 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: 0.0001
- 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
- num_epochs: 10
### 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 |
|:-------------:|:-----:|:----:|:---------------:|:-------------:|:----------:|:------:|:----------:|:---------------------:|:------------------:|:--------------:|:------------------:|:---------------------:|:------------------:|:--------------:|:------------------:|:-------------------:|:----------------:|:------------:|:----------------:|:-----------------:|:--------------:|:----------:|:----------------:|
| No log | 1.0 | 45 | 0.6876 | 0.0238 | 0.0135 | 0.0172 | 74 | 0.0 | 0.0 | 0.0 | 72 | 0.0 | 0.0 | 0.0 | 49 | 0.0104 | 0.0319 | 0.0156 | 596 | 0.0105 | 0.0253 | 0.0149 | 0.7882 |
| No log | 2.0 | 90 | 0.4180 | 0.1235 | 0.1351 | 0.1290 | 74 | 0.25 | 0.1389 | 0.1786 | 72 | 0.0 | 0.0 | 0.0 | 49 | 0.0989 | 0.1846 | 0.1288 | 596 | 0.1052 | 0.1643 | 0.1283 | 0.8636 |
| No log | 3.0 | 135 | 0.3403 | 0.3647 | 0.4189 | 0.3899 | 74 | 0.3803 | 0.375 | 0.3776 | 72 | 0.0968 | 0.0612 | 0.075 | 49 | 0.1694 | 0.2903 | 0.2140 | 596 | 0.1937 | 0.2958 | 0.2341 | 0.8853 |
| No log | 4.0 | 180 | 0.3404 | 0.4 | 0.4054 | 0.4027 | 74 | 0.3721 | 0.4444 | 0.4051 | 72 | 0.1471 | 0.2041 | 0.1709 | 49 | 0.2298 | 0.3339 | 0.2722 | 596 | 0.2475 | 0.3426 | 0.2874 | 0.8974 |
| No log | 5.0 | 225 | 0.3894 | 0.4545 | 0.5405 | 0.4938 | 74 | 0.4022 | 0.5139 | 0.4512 | 72 | 0.1587 | 0.2041 | 0.1786 | 49 | 0.2725 | 0.3859 | 0.3194 | 596 | 0.2916 | 0.4008 | 0.3376 | 0.9016 |
| No log | 6.0 | 270 | 0.3922 | 0.4565 | 0.5676 | 0.5060 | 74 | 0.4118 | 0.4861 | 0.4459 | 72 | 0.2051 | 0.1633 | 0.1818 | 49 | 0.3081 | 0.4849 | 0.3768 | 596 | 0.3241 | 0.4728 | 0.3846 | 0.9070 |
| No log | 7.0 | 315 | 0.4751 | 0.4819 | 0.5405 | 0.5096 | 74 | 0.3953 | 0.4722 | 0.4304 | 72 | 0.1525 | 0.1837 | 0.1667 | 49 | 0.2579 | 0.3708 | 0.3042 | 596 | 0.2802 | 0.3843 | 0.3241 | 0.9011 |
| No log | 8.0 | 360 | 0.4684 | 0.4257 | 0.5811 | 0.4914 | 74 | 0.4 | 0.5278 | 0.4551 | 72 | 0.1429 | 0.2041 | 0.1681 | 49 | 0.3025 | 0.4446 | 0.3601 | 596 | 0.3117 | 0.4501 | 0.3683 | 0.9063 |
| No log | 9.0 | 405 | 0.5157 | 0.4554 | 0.6216 | 0.5257 | 74 | 0.4211 | 0.5556 | 0.4790 | 72 | 0.1515 | 0.2041 | 0.1739 | 49 | 0.2969 | 0.4329 | 0.3522 | 596 | 0.3130 | 0.4475 | 0.3684 | 0.9071 |
| No log | 10.0 | 450 | 0.5566 | 0.4434 | 0.6351 | 0.5222 | 74 | 0.3663 | 0.5139 | 0.4277 | 72 | 0.1507 | 0.2245 | 0.1803 | 49 | 0.3132 | 0.4430 | 0.3669 | 596 | 0.3197 | 0.4539 | 0.3751 | 0.9062 |
### Framework versions
- PEFT 0.12.0
- Transformers 4.44.0
- Pytorch 2.4.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1 |