Hugging Face's logo

cite-text-analysis
/
case-analysis-distilbert-base-cased

Text Classification
Transformers
TensorBoard
Safetensors
distilbert
Generated from Trainer
Model card Files Metrics Community
case-analysis-distilbert-base-cased / README.md
cite-text-analysis's picture
cite-text-analysis
Upload README.md with huggingface_hub
977f570 verified
preview code
|
raw
history blame contribute delete
10.5 kB
metadata
license: apache-2.0
tags:
  - generated_from_trainer
base_model: distilbert/distilbert-base-cased
metrics:
  - accuracy
  - precision
  - recall
model-index:
  - name: case-analysis-distilbert-base-cased
    results: []

Metrics

  • loss: 1.8402
  • accuracy: 0.8085
  • precision: 0.7983
  • recall: 0.8085
  • precision_macro: 0.6608
  • recall_macro: 0.6429
  • macro_fpr: 0.0935
  • weighted_fpr: 0.0732
  • weighted_specificity: 0.8548
  • macro_specificity: 0.9158
  • weighted_sensitivity: 0.8085
  • macro_sensitivity: 0.6429
  • f1_micro: 0.8085
  • f1_macro: 0.6478
  • f1_weighted: 0.8018
  • runtime: 131.6318
  • samples_per_second: 3.4110
  • steps_per_second: 0.4330

case-analysis-distilbert-base-cased

This model is a fine-tuned version of distilbert/distilbert-base-cased on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 1.8402
  • Accuracy: 0.8085
  • Precision: 0.7983
  • Recall: 0.8085
  • Precision Macro: 0.6461
  • Recall Macro: 0.6218
  • Macro Fpr: 0.0984
  • Weighted Fpr: 0.0771
  • Weighted Specificity: 0.8479
  • Macro Specificity: 0.9119
  • Weighted Sensitivity: 0.7996
  • Macro Sensitivity: 0.6218
  • F1 Micro: 0.7996
  • F1 Macro: 0.6245
  • F1 Weighted: 0.7887

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: 8
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 30
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Accuracy Precision Recall Precision Macro Recall Macro Macro Fpr Weighted Fpr Weighted Specificity Macro Specificity Weighted Sensitivity Macro Sensitivity F1 Micro F1 Macro F1 Weighted
No log 1.0 224 0.7001 0.7661 0.7311 0.7661 0.5791 0.5137 0.1330 0.0923 0.7614 0.8819 0.7661 0.5137 0.7661 0.5270 0.7333
No log 2.0 448 0.7388 0.7751 0.7315 0.7751 0.5585 0.5464 0.1208 0.0882 0.7908 0.8915 0.7751 0.5464 0.7751 0.5487 0.7493
0.7066 3.0 672 0.7229 0.8018 0.7605 0.8018 0.5932 0.5708 0.1076 0.0761 0.8090 0.9027 0.8018 0.5708 0.8018 0.5767 0.7760
0.7066 4.0 896 0.8331 0.8062 0.7896 0.8062 0.6675 0.6115 0.1018 0.0742 0.8218 0.9070 0.8062 0.6115 0.8062 0.6301 0.7934
0.3654 5.0 1120 1.2300 0.7684 0.7699 0.7684 0.6085 0.6131 0.1066 0.0913 0.8542 0.9056 0.7684 0.6131 0.7684 0.5896 0.7611
0.3654 6.0 1344 1.0698 0.8129 0.7940 0.8129 0.6864 0.6153 0.0957 0.0712 0.8406 0.9134 0.8129 0.6153 0.8129 0.6300 0.7972
0.2047 7.0 1568 1.3300 0.7884 0.7960 0.7884 0.6412 0.5959 0.1044 0.0821 0.8421 0.9076 0.7884 0.5959 0.7884 0.6141 0.7892
0.2047 8.0 1792 1.3870 0.8107 0.7861 0.8107 0.6467 0.6063 0.0983 0.0722 0.8318 0.9106 0.8107 0.6063 0.8107 0.6163 0.7947
0.0795 9.0 2016 1.5031 0.7951 0.7719 0.7951 0.6275 0.5969 0.1040 0.0791 0.8320 0.9068 0.7951 0.5969 0.7951 0.6036 0.7803
0.0795 10.0 2240 1.6304 0.7728 0.7796 0.7728 0.6171 0.6233 0.1060 0.0892 0.8561 0.9072 0.7728 0.6233 0.7728 0.6196 0.7759
0.0795 11.0 2464 1.6553 0.8040 0.7802 0.8040 0.6405 0.6047 0.1003 0.0751 0.8333 0.9093 0.8040 0.6047 0.8040 0.6097 0.7884
0.0309 12.0 2688 1.6668 0.7996 0.7776 0.7996 0.6247 0.6084 0.0999 0.0771 0.8431 0.9107 0.7996 0.6084 0.7996 0.6073 0.7861
0.0309 13.0 2912 1.7548 0.8040 0.7724 0.8040 0.6059 0.5847 0.1030 0.0751 0.8216 0.9064 0.8040 0.5847 0.8040 0.5912 0.7846
0.0225 14.0 3136 1.6691 0.8107 0.7736 0.8107 0.5965 0.6044 0.0974 0.0722 0.8336 0.9111 0.8107 0.6044 0.8107 0.5998 0.7909
0.0225 15.0 3360 1.8751 0.8040 0.7897 0.8040 0.6516 0.6081 0.1007 0.0751 0.8322 0.9091 0.8040 0.6081 0.8040 0.6251 0.7939
0.0048 16.0 3584 1.8402 0.8085 0.7983 0.8085 0.6608 0.6429 0.0935 0.0732 0.8548 0.9158 0.8085 0.6429 0.8085 0.6478 0.8018
0.0048 17.0 3808 1.9124 0.7951 0.7871 0.7951 0.6331 0.6237 0.1001 0.0791 0.8456 0.9102 0.7951 0.6237 0.7951 0.6250 0.7891
0.0069 18.0 4032 1.8857 0.7973 0.7794 0.7973 0.6268 0.5972 0.1048 0.0781 0.8240 0.9053 0.7973 0.5972 0.7973 0.6062 0.7847
0.0069 19.0 4256 1.9492 0.8062 0.7813 0.8062 0.6467 0.6015 0.1006 0.0742 0.8281 0.9086 0.8062 0.6015 0.8062 0.6107 0.7895
0.0069 20.0 4480 1.8994 0.8085 0.7849 0.8085 0.6417 0.6067 0.0988 0.0732 0.8322 0.9102 0.8085 0.6067 0.8085 0.6144 0.7932
0.0034 21.0 4704 1.9819 0.8040 0.7898 0.8040 0.6748 0.6325 0.0976 0.0751 0.8439 0.9120 0.8040 0.6325 0.8040 0.6429 0.7942
0.0034 22.0 4928 2.0181 0.8062 0.7880 0.8062 0.6736 0.6204 0.0977 0.0742 0.8408 0.9118 0.8062 0.6204 0.8062 0.6293 0.7930
0.0001 23.0 5152 2.0305 0.8062 0.7880 0.8062 0.6736 0.6204 0.0977 0.0742 0.8408 0.9118 0.8062 0.6204 0.8062 0.6293 0.7930
0.0001 24.0 5376 2.0249 0.8040 0.7801 0.8040 0.6448 0.6004 0.1019 0.0751 0.8256 0.9074 0.8040 0.6004 0.8040 0.6092 0.7877
0.0 25.0 5600 2.0139 0.8018 0.7848 0.8018 0.6514 0.6226 0.0984 0.0761 0.8438 0.9114 0.8018 0.6226 0.8018 0.6272 0.7908
0.0 26.0 5824 2.0075 0.8040 0.7868 0.8040 0.6586 0.6281 0.0961 0.0751 0.8487 0.9132 0.8040 0.6281 0.8040 0.6305 0.7926
0.0026 27.0 6048 2.0155 0.8040 0.7868 0.8040 0.6586 0.6281 0.0961 0.0751 0.8487 0.9132 0.8040 0.6281 0.8040 0.6305 0.7926
0.0026 28.0 6272 2.0191 0.8040 0.7865 0.8040 0.6586 0.6237 0.0970 0.0751 0.8463 0.9126 0.8040 0.6237 0.8040 0.6283 0.7923
0.0026 29.0 6496 2.0225 0.8040 0.7865 0.8040 0.6586 0.6237 0.0970 0.0751 0.8463 0.9126 0.8040 0.6237 0.8040 0.6283 0.7923
0.0 30.0 6720 2.0343 0.7996 0.7821 0.7996 0.6461 0.6218 0.0984 0.0771 0.8479 0.9119 0.7996 0.6218 0.7996 0.6245 0.7887

Framework versions

  • Transformers 4.40.2
  • Pytorch 2.2.1+cu121
  • Datasets 2.19.1
  • Tokenizers 0.19.1