haryoaw's picture
Initial Commit
1262aec 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_1333
    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: 0.804343713345189
          - name: F1
            type: f1
            value: 0.7729337846791551

scenario-NON-KD-SCR-D2_data-AmazonScience_massive_all_1_1333

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: 1.9549
  • Accuracy: 0.8043
  • F1: 0.7729

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: 333
  • 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
1.5003 0.27 5000 1.4803 0.6001 0.4750
1.0696 0.53 10000 1.0629 0.7133 0.6331
0.8993 0.8 15000 0.9337 0.7549 0.6941
0.6223 1.07 20000 0.9072 0.7691 0.7180
0.6122 1.34 25000 0.8510 0.7820 0.7315
0.5894 1.6 30000 0.8077 0.7893 0.7424
0.5692 1.87 35000 0.7765 0.7970 0.7542
0.3814 2.14 40000 0.8363 0.7967 0.7574
0.3888 2.41 45000 0.8421 0.7970 0.7587
0.3886 2.67 50000 0.8302 0.8035 0.7672
0.407 2.94 55000 0.8024 0.8052 0.7692
0.2541 3.21 60000 0.8741 0.8075 0.7743
0.2687 3.47 65000 0.8732 0.8077 0.7736
0.2857 3.74 70000 0.8756 0.8060 0.7696
0.2347 4.01 75000 0.9543 0.8079 0.7698
0.1788 4.28 80000 1.0152 0.8046 0.7722
0.1896 4.54 85000 0.9846 0.8074 0.7737
0.207 4.81 90000 0.9621 0.8066 0.7709
0.1175 5.08 95000 1.1112 0.8069 0.7770
0.1389 5.34 100000 1.0817 0.8085 0.7697
0.145 5.61 105000 1.0714 0.8048 0.7755
0.1591 5.88 110000 1.0711 0.8051 0.7788
0.1008 6.15 115000 1.2102 0.8086 0.7794
0.1145 6.41 120000 1.2193 0.8066 0.7680
0.1268 6.68 125000 1.1596 0.8066 0.7787
0.1238 6.95 130000 1.1909 0.8084 0.7752
0.0894 7.22 135000 1.3205 0.8054 0.7792
0.0983 7.48 140000 1.3175 0.8085 0.7742
0.1056 7.75 145000 1.3009 0.8034 0.7725
0.0796 8.02 150000 1.3048 0.8076 0.7797
0.079 8.28 155000 1.3814 0.8031 0.7787
0.088 8.55 160000 1.3723 0.8077 0.7798
0.0936 8.82 165000 1.3706 0.8052 0.7770
0.0558 9.09 170000 1.4600 0.8057 0.7741
0.0655 9.35 175000 1.4752 0.8023 0.7737
0.0752 9.62 180000 1.4717 0.8032 0.7726
0.0799 9.89 185000 1.4064 0.8071 0.7783
0.0536 10.15 190000 1.5280 0.8029 0.7703
0.0596 10.42 195000 1.5051 0.8045 0.7772
0.0605 10.69 200000 1.5007 0.8040 0.7771
0.0723 10.96 205000 1.5009 0.8053 0.7776
0.0424 11.22 210000 1.6065 0.7998 0.7671
0.0587 11.49 215000 1.5795 0.8014 0.7664
0.06 11.76 220000 1.6210 0.7959 0.7637
0.0444 12.03 225000 1.6069 0.8042 0.7778
0.0427 12.29 230000 1.5564 0.8031 0.7745
0.0462 12.56 235000 1.6148 0.8007 0.7716
0.0507 12.83 240000 1.6110 0.8022 0.7715
0.0398 13.09 245000 1.6613 0.8036 0.7728
0.0411 13.36 250000 1.6634 0.8047 0.7782
0.0462 13.63 255000 1.6522 0.8046 0.7756
0.0489 13.9 260000 1.6642 0.8015 0.7738
0.0395 14.16 265000 1.6743 0.8019 0.7683
0.0407 14.43 270000 1.6974 0.8036 0.7713
0.0435 14.7 275000 1.6568 0.8038 0.7725
0.0383 14.96 280000 1.6867 0.8044 0.7736
0.0331 15.23 285000 1.7553 0.8015 0.7720
0.0355 15.5 290000 1.7463 0.7982 0.7667
0.0309 15.77 295000 1.7347 0.8017 0.7735
0.0205 16.03 300000 1.7513 0.8028 0.7714
0.0322 16.3 305000 1.7507 0.8013 0.7726
0.031 16.57 310000 1.7373 0.8043 0.7724
0.0352 16.84 315000 1.7256 0.8022 0.7706
0.0246 17.1 320000 1.7548 0.8036 0.7712
0.0276 17.37 325000 1.8002 0.7984 0.7686
0.0259 17.64 330000 1.7736 0.8011 0.7701
0.0218 17.9 335000 1.8022 0.7996 0.7718
0.0228 18.17 340000 1.8162 0.8021 0.7671
0.0202 18.44 345000 1.8392 0.8011 0.7678
0.0251 18.71 350000 1.7928 0.7997 0.7673
0.0263 18.97 355000 1.8359 0.8000 0.7671
0.0202 19.24 360000 1.8644 0.8021 0.7739
0.0215 19.51 365000 1.8412 0.7988 0.7685
0.0277 19.77 370000 1.7734 0.8009 0.7685
0.018 20.04 375000 1.8197 0.7997 0.7716
0.0188 20.31 380000 1.8411 0.8017 0.7703
0.0233 20.58 385000 1.8631 0.7993 0.7658
0.0216 20.84 390000 1.8590 0.8010 0.7676
0.0159 21.11 395000 1.8778 0.7977 0.7639
0.0191 21.38 400000 1.8380 0.8028 0.7732
0.0173 21.65 405000 1.8699 0.8027 0.7754
0.0216 21.91 410000 1.9152 0.7983 0.7651
0.0139 22.18 415000 1.8876 0.7987 0.7680
0.0172 22.45 420000 1.8977 0.7996 0.7656
0.0135 22.71 425000 1.8767 0.7996 0.7687
0.016 22.98 430000 1.8973 0.8015 0.7689
0.0135 23.25 435000 1.9112 0.8007 0.7668
0.0162 23.52 440000 1.9312 0.7994 0.7695
0.0178 23.78 445000 1.8928 0.8012 0.7679
0.0089 24.05 450000 1.9162 0.8012 0.7643
0.0123 24.32 455000 1.9334 0.8006 0.7671
0.0137 24.58 460000 1.8863 0.8018 0.7683
0.01 24.85 465000 1.9100 0.8012 0.7667
0.008 25.12 470000 1.9313 0.8011 0.7687
0.0113 25.39 475000 1.9219 0.8025 0.7718
0.0129 25.65 480000 1.9310 0.8015 0.7705
0.0165 25.92 485000 1.9018 0.8033 0.7736
0.0085 26.19 490000 1.9537 0.8035 0.7705
0.0108 26.46 495000 1.9116 0.8019 0.7701
0.0086 26.72 500000 1.9266 0.8034 0.7720
0.0094 26.99 505000 1.9232 0.8036 0.7721
0.0083 27.26 510000 1.9425 0.8027 0.7728
0.0078 27.52 515000 1.9477 0.8044 0.7727
0.0095 27.79 520000 1.9127 0.8031 0.7687
0.0067 28.06 525000 1.9433 0.8030 0.7692
0.009 28.33 530000 1.9325 0.8042 0.7719
0.0074 28.59 535000 1.9332 0.8048 0.7740
0.0071 28.86 540000 1.9393 0.8049 0.7734
0.006 29.13 545000 1.9489 0.8039 0.7719
0.0061 29.39 550000 1.9571 0.8041 0.7716
0.0065 29.66 555000 1.9528 0.8046 0.7728
0.0084 29.93 560000 1.9549 0.8043 0.7729

Framework versions

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