Hugging Face's logo

tthhanh
/
legal_qwen

PEFT
Safetensors
Generated from Trainer
Model card Files Community
legal_qwen / README.md
tthhanh's picture
tthhanh
End of training
64f60c4 verified
preview code
|
raw
history blame
6.59 kB
metadata
base_model: Qwen/Qwen2-7B
library_name: peft
license: apache-2.0
tags:
  - generated_from_trainer
model-index:
  - name: legal_qwen
    results: []

legal_qwen

This model is a fine-tuned version of 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