Hugging Face's logo

hcene
/
CamemBERT_xnli_fintuned_on_legal_data

Text Classification
Transformers
TensorBoard
Safetensors
camembert
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
CamemBERT_xnli_fintuned_on_legal_data / README.md
hcene's picture
hcene
hcene/Camembert-xnli
a278bca verified
preview code
|
raw
history blame contribute delete
6.96 kB
metadata
base_model: mtheo/camembert-base-xnli
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - precision
  - recall
  - f1
model-index:
  - name: legal-data-mDEBERTa-V3-base-mnli-xnli
    results: []

legal-data-mDEBERTa-V3-base-mnli-xnli

This model is a fine-tuned version of mtheo/camembert-base-xnli on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.7972
  • Accuracy: 0.7706
  • Precision: 0.7891
  • Recall: 0.7711
  • F1: 0.7697
  • Ratio: 0.3154

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: 2e-05
  • train_batch_size: 20
  • eval_batch_size: 20
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.1
  • lr_scheduler_warmup_steps: 4
  • num_epochs: 15
  • label_smoothing_factor: 0.1

Training results

Training Loss Epoch Step Validation Loss Accuracy Precision Recall F1 Ratio
0.9937 0.17 10 0.8509 0.6093 0.7022 0.6086 0.6168 0.1935
0.8433 0.34 20 0.7131 0.6667 0.6703 0.6687 0.6666 0.3262
0.8683 0.52 30 0.7101 0.7312 0.7350 0.7324 0.7324 0.3262
0.8536 0.69 40 0.7542 0.6989 0.7152 0.6999 0.7053 0.3011
0.7964 0.86 50 0.7249 0.7670 0.7773 0.7677 0.7630 0.3297
0.7651 1.03 60 0.8253 0.7455 0.7517 0.7465 0.7450 0.3262
0.7658 1.21 70 0.8282 0.6953 0.7366 0.6956 0.7030 0.2688
0.7297 1.38 80 0.8694 0.7634 0.7771 0.7641 0.7612 0.3226
0.7726 1.55 90 0.7898 0.7097 0.7112 0.7112 0.7111 0.3297
0.7107 1.72 100 0.8279 0.7599 0.7642 0.7608 0.7589 0.3297
0.7204 1.9 110 0.9353 0.7240 0.7728 0.7240 0.7279 0.2724
0.7241 2.07 120 0.7903 0.7527 0.7818 0.7530 0.7531 0.3011
0.6683 2.24 130 0.8139 0.7384 0.7931 0.7382 0.7397 0.2760
0.6832 2.41 140 0.8339 0.7993 0.8268 0.7996 0.7913 0.3297
0.7397 2.59 150 0.8309 0.7849 0.7997 0.7855 0.7801 0.3297
0.6945 2.76 160 0.7860 0.7240 0.7259 0.7253 0.7247 0.3297
0.7067 2.93 170 0.7046 0.7957 0.8152 0.7962 0.7898 0.3297
0.6759 3.1 180 0.7465 0.7634 0.7703 0.7643 0.7611 0.3297
0.6673 3.28 190 0.8461 0.8029 0.8234 0.8033 0.7972 0.3297
0.6748 3.45 200 0.8701 0.8065 0.8354 0.8068 0.7987 0.3297
0.7638 3.62 210 0.7501 0.8136 0.8521 0.8138 0.8043 0.3297
0.6426 3.79 220 0.7165 0.8100 0.8379 0.8103 0.8029 0.3297
0.6569 3.97 230 0.7244 0.8100 0.8415 0.8103 0.8020 0.3297
0.7068 4.14 240 0.7448 0.8100 0.8499 0.8102 0.8000 0.3297
0.6544 4.31 250 0.8241 0.8065 0.8476 0.8066 0.7957 0.3297
0.6261 4.48 260 0.8409 0.7634 0.7692 0.7643 0.7617 0.3297
0.6428 4.66 270 0.7887 0.7491 0.7528 0.7501 0.7484 0.3297
0.657 4.83 280 0.7534 0.7706 0.7791 0.7714 0.7677 0.3297
0.6895 5.0 290 0.7863 0.8100 0.8379 0.8103 0.8029 0.3297
0.6422 5.17 300 0.7908 0.8136 0.8439 0.8139 0.8062 0.3297
0.5933 5.34 310 0.8330 0.7885 0.8031 0.7891 0.7869 0.3226
0.5863 5.52 320 0.8494 0.7527 0.7598 0.7536 0.7535 0.3226
0.6787 5.69 330 0.7748 0.7742 0.7823 0.7750 0.7716 0.3297
0.6761 5.86 340 0.7256 0.7814 0.7929 0.7820 0.7775 0.3297
0.6974 6.03 350 0.7711 0.8029 0.8208 0.8033 0.7980 0.3297
0.6083 6.21 360 0.8435 0.7993 0.8191 0.7997 0.7951 0.3262
0.5897 6.38 370 0.8773 0.7849 0.8124 0.7852 0.7831 0.3118
0.6076 6.55 380 0.8255 0.7634 0.7683 0.7644 0.7623 0.3297
0.6709 6.72 390 0.7865 0.7527 0.7551 0.7538 0.7530 0.3297
0.7063 6.9 400 0.7898 0.8029 0.8234 0.8033 0.7972 0.3297
0.6804 7.07 410 0.7804 0.7921 0.8152 0.7925 0.7848 0.3297
0.6227 7.24 420 0.7515 0.7706 0.7778 0.7714 0.7683 0.3297
0.6482 7.41 430 0.7758 0.7670 0.7725 0.7679 0.7656 0.3297
0.6072 7.59 440 0.8077 0.7706 0.7792 0.7714 0.7693 0.3262
0.5695 7.76 450 0.8460 0.7921 0.8068 0.7926 0.7892 0.3262
0.6097 7.93 460 0.7856 0.7993 0.8191 0.7997 0.7951 0.3262
0.5591 8.1 470 0.7812 0.8136 0.8447 0.8138 0.8074 0.3262
0.5573 8.28 480 0.7249 0.7849 0.7930 0.7857 0.7828 0.3297
0.6128 8.45 490 0.7245 0.7921 0.8006 0.7928 0.7901 0.3297
0.6072 8.62 500 0.7732 0.7885 0.8038 0.7891 0.7852 0.3262
0.6276 8.79 510 0.8017 0.7885 0.8038 0.7891 0.7852 0.3262
0.647 8.97 520 0.7998 0.7885 0.8038 0.7891 0.7852 0.3262
0.5924 9.14 530 0.7972 0.7706 0.7891 0.7711 0.7697 0.3154

Framework versions

  • Transformers 4.39.0.dev0
  • Pytorch 2.1.0+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2