haryoaw's picture
Initial Commit
53ac4a2 verified
metadata
license: mit
base_model: xlm-roberta-base
tags:
  - generated_from_trainer
datasets:
  - massive
metrics:
  - accuracy
  - f1
model-index:
  - name: scenario-NON-KD-SCR-D2_data-AmazonScience_massive_all_1_1111
    results:
      - task:
          name: Text Classification
          type: text-classification
        dataset:
          name: massive
          type: massive
          config: all_1.1
          split: validation
          args: all_1.1
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.8041450679178176
          - name: F1
            type: f1
            value: 0.7741242752130463

scenario-NON-KD-SCR-D2_data-AmazonScience_massive_all_1_1111

This model is a fine-tuned version of xlm-roberta-base on the massive dataset. It achieves the following results on the evaluation set:

  • Loss: 1.9383
  • Accuracy: 0.8041
  • F1: 0.7741

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: 32
  • eval_batch_size: 32
  • seed: 111
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 30

Training results

Training Loss Epoch Step Validation Loss Accuracy F1
1.625 0.27 5000 1.5556 0.5832 0.4573
1.1017 0.53 10000 1.0835 0.7113 0.6333
0.9149 0.8 15000 0.9682 0.7414 0.6802
0.652 1.07 20000 0.8972 0.7697 0.7151
0.6198 1.34 25000 0.8545 0.7780 0.7314
0.6266 1.6 30000 0.8015 0.7901 0.7412
0.5837 1.87 35000 0.7959 0.7950 0.7590
0.3694 2.14 40000 0.8423 0.7980 0.7569
0.3908 2.41 45000 0.8553 0.7916 0.7578
0.4034 2.67 50000 0.7928 0.8039 0.7667
0.391 2.94 55000 0.8013 0.8061 0.7698
0.2482 3.21 60000 0.8858 0.8043 0.7719
0.2658 3.47 65000 0.9176 0.8043 0.7679
0.2786 3.74 70000 0.8828 0.8052 0.7720
0.2445 4.01 75000 0.9493 0.8075 0.7717
0.1873 4.28 80000 1.0095 0.8028 0.7741
0.2023 4.54 85000 0.9853 0.8041 0.7703
0.2141 4.81 90000 0.9830 0.8081 0.7797
0.1302 5.08 95000 1.0565 0.8081 0.7790
0.1478 5.34 100000 1.0957 0.8061 0.7770
0.1544 5.61 105000 1.0758 0.8036 0.7636
0.1529 5.88 110000 1.1086 0.8057 0.7671
0.1056 6.15 115000 1.2581 0.8024 0.7727
0.1119 6.41 120000 1.2179 0.8046 0.7757
0.1191 6.68 125000 1.2204 0.8039 0.7752
0.1276 6.95 130000 1.1614 0.8056 0.7735
0.087 7.22 135000 1.3199 0.8037 0.7744
0.0974 7.48 140000 1.3204 0.8062 0.7765
0.0977 7.75 145000 1.3529 0.7993 0.7702
0.0741 8.02 150000 1.3510 0.8030 0.7783
0.0736 8.28 155000 1.4104 0.8029 0.7730
0.0836 8.55 160000 1.4256 0.8000 0.7687
0.0874 8.82 165000 1.4120 0.8024 0.7680
0.06 9.09 170000 1.4139 0.8029 0.7740
0.0722 9.35 175000 1.4836 0.8010 0.7712
0.0729 9.62 180000 1.4753 0.8003 0.7724
0.0733 9.89 185000 1.4762 0.8015 0.7747
0.0537 10.15 190000 1.5196 0.8022 0.7735
0.0579 10.42 195000 1.5303 0.8022 0.7741
0.0633 10.69 200000 1.5843 0.8023 0.7730
0.0702 10.96 205000 1.5198 0.8042 0.7779
0.0473 11.22 210000 1.6088 0.8008 0.7716
0.0506 11.49 215000 1.6281 0.8001 0.7740
0.0601 11.76 220000 1.5632 0.8032 0.7758
0.0369 12.03 225000 1.6079 0.8012 0.7699
0.0492 12.29 230000 1.6162 0.8004 0.7726
0.0473 12.56 235000 1.6604 0.8004 0.7725
0.0449 12.83 240000 1.5631 0.8023 0.7763
0.0334 13.09 245000 1.6218 0.8049 0.7734
0.0426 13.36 250000 1.6875 0.8004 0.7736
0.0483 13.63 255000 1.6627 0.8028 0.7736
0.0514 13.9 260000 1.6705 0.8014 0.7705
0.0357 14.16 265000 1.7121 0.8008 0.7759
0.0313 14.43 270000 1.7074 0.7993 0.7714
0.0405 14.7 275000 1.6907 0.7973 0.7619
0.0473 14.96 280000 1.7018 0.8006 0.7707
0.0364 15.23 285000 1.7487 0.8009 0.7725
0.0369 15.5 290000 1.7177 0.7996 0.7635
0.0407 15.77 295000 1.7514 0.7981 0.7676
0.0215 16.03 300000 1.8013 0.8003 0.7738
0.0251 16.3 305000 1.7813 0.8001 0.7700
0.0306 16.57 310000 1.7511 0.8029 0.7748
0.0328 16.84 315000 1.7910 0.8015 0.7750
0.0191 17.1 320000 1.8131 0.8002 0.7663
0.0231 17.37 325000 1.7831 0.8027 0.7771
0.0274 17.64 330000 1.7864 0.8025 0.7743
0.0355 17.9 335000 1.8057 0.8004 0.7693
0.019 18.17 340000 1.8307 0.8001 0.7704
0.0255 18.44 345000 1.8017 0.7999 0.7681
0.033 18.71 350000 1.8074 0.7983 0.7701
0.0329 18.97 355000 1.8416 0.7988 0.7690
0.0216 19.24 360000 1.8396 0.8003 0.7719
0.0234 19.51 365000 1.8631 0.7999 0.7707
0.0228 19.77 370000 1.8195 0.8031 0.7751
0.0148 20.04 375000 1.8301 0.8026 0.7731
0.0203 20.31 380000 1.8525 0.8009 0.7709
0.0183 20.58 385000 1.8466 0.7978 0.7678
0.0171 20.84 390000 1.8859 0.8016 0.7751
0.0156 21.11 395000 1.8790 0.8000 0.7698
0.0169 21.38 400000 1.8781 0.8015 0.7733
0.0193 21.65 405000 1.8454 0.8016 0.7723
0.0157 21.91 410000 1.8695 0.8008 0.7710
0.0111 22.18 415000 1.8899 0.8010 0.7718
0.0178 22.45 420000 1.8696 0.7990 0.7692
0.0183 22.71 425000 1.8613 0.8006 0.7722
0.0202 22.98 430000 1.8738 0.7991 0.7685
0.0127 23.25 435000 1.8803 0.8039 0.7753
0.0139 23.52 440000 1.9212 0.7983 0.7669
0.0149 23.78 445000 1.8538 0.8016 0.7716
0.0094 24.05 450000 1.9183 0.8010 0.7729
0.0125 24.32 455000 1.9316 0.7997 0.7709
0.015 24.58 460000 1.8689 0.8011 0.7713
0.013 24.85 465000 1.9028 0.8015 0.7734
0.0114 25.12 470000 1.9559 0.8003 0.7705
0.0105 25.39 475000 1.9195 0.8013 0.7705
0.0138 25.65 480000 1.8951 0.8032 0.7739
0.0125 25.92 485000 1.9088 0.8024 0.7735
0.0087 26.19 490000 1.9183 0.8001 0.7687
0.0089 26.46 495000 1.9353 0.8027 0.7754
0.0088 26.72 500000 1.8883 0.8005 0.7692
0.0114 26.99 505000 1.9141 0.8035 0.7741
0.0094 27.26 510000 1.9412 0.8032 0.7730
0.007 27.52 515000 1.9465 0.8029 0.7735
0.0077 27.79 520000 1.9341 0.8046 0.7765
0.0065 28.06 525000 1.9372 0.8038 0.7739
0.0092 28.33 530000 1.9510 0.8030 0.7740
0.0103 28.59 535000 1.9216 0.8033 0.7724
0.0082 28.86 540000 1.9275 0.8031 0.7726
0.0076 29.13 545000 1.9477 0.8036 0.7735
0.0063 29.39 550000 1.9361 0.8033 0.7722
0.0073 29.66 555000 1.9427 0.8038 0.7737
0.006 29.93 560000 1.9383 0.8041 0.7741

Framework versions

  • Transformers 4.33.3
  • Pytorch 2.1.1+cu121
  • Datasets 2.14.5
  • Tokenizers 0.13.3