csNoHug's picture
Training complete
0a0dd10
metadata
license: apache-2.0
base_model: albert-base-v2
tags:
  - generated_from_trainer
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: albert-base-v2-finetuned-ner-cadec-no-iob
    results: []

albert-base-v2-finetuned-ner-cadec-no-iob

This model is a fine-tuned version of albert-base-v2 on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.5037
  • Precision: 0.5849
  • Recall: 0.6227
  • F1: 0.6032
  • Accuracy: 0.9311
  • Adr Precision: 0.5065
  • Adr Recall: 0.5608
  • Adr F1: 0.5323
  • Disease Precision: 0.52
  • Disease Recall: 0.4062
  • Disease F1: 0.4561
  • Drug Precision: 0.9121
  • Drug Recall: 0.9222
  • Drug F1: 0.9171
  • Finding Precision: 0.1875
  • Finding Recall: 0.1875
  • Finding F1: 0.1875
  • Symptom Precision: 0.4839
  • Symptom Recall: 0.5172
  • Symptom F1: 0.5000
  • Macro Avg F1: 0.5186
  • Weighted Avg F1: 0.6047

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: 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: 35

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Accuracy Adr Precision Adr Recall Adr F1 Disease Precision Disease Recall Disease F1 Drug Precision Drug Recall Drug F1 Finding Precision Finding Recall Finding F1 Symptom Precision Symptom Recall Symptom F1 Macro Avg F1 Weighted Avg F1
No log 1.0 125 0.2244 0.5211 0.6029 0.5590 0.9215 0.4547 0.6103 0.5211 0.3864 0.5312 0.4474 0.8276 0.8 0.8136 0.0 0.0 0.0 0.0 0.0 0.0 0.3564 0.5455
No log 2.0 250 0.2082 0.5448 0.5937 0.5682 0.9240 0.4700 0.5649 0.5131 0.4348 0.3125 0.3636 0.8722 0.8722 0.8722 0.1892 0.2188 0.2029 0.6667 0.0690 0.125 0.4154 0.5641
No log 3.0 375 0.2113 0.5416 0.6016 0.5700 0.9273 0.4863 0.5505 0.5164 0.48 0.375 0.4211 0.8182 0.85 0.8338 0.1622 0.1875 0.1739 0.4091 0.6207 0.4932 0.4877 0.5724
0.187 4.0 500 0.2257 0.5418 0.6069 0.5725 0.9281 0.4739 0.5423 0.5058 0.3548 0.3438 0.3492 0.8944 0.8944 0.8944 0.1607 0.2812 0.2045 0.5926 0.5517 0.5714 0.5051 0.5813
0.187 5.0 625 0.2483 0.5788 0.6253 0.6011 0.9283 0.4957 0.5918 0.5395 0.6111 0.3438 0.4400 0.8883 0.8833 0.8858 0.2083 0.1562 0.1786 0.6316 0.4138 0.5 0.5088 0.6008
0.187 6.0 750 0.2584 0.5572 0.6042 0.5797 0.9242 0.4843 0.5423 0.5117 0.4 0.375 0.3871 0.8989 0.8889 0.8939 0.1951 0.25 0.2192 0.5 0.5172 0.5085 0.5041 0.5847
0.187 7.0 875 0.2676 0.5640 0.5989 0.5809 0.9261 0.4836 0.5464 0.5131 1.0 0.0938 0.1714 0.9096 0.8944 0.9020 0.1964 0.3438 0.25 0.6667 0.4828 0.56 0.4793 0.5817
0.0608 8.0 1000 0.2623 0.5797 0.6187 0.5986 0.9335 0.5121 0.5670 0.5382 0.5556 0.3125 0.4000 0.8944 0.8944 0.8944 0.1562 0.1562 0.1562 0.4286 0.6207 0.5070 0.4992 0.5996
0.0608 9.0 1125 0.2968 0.5754 0.6293 0.6011 0.9314 0.5162 0.5897 0.5505 0.4062 0.4062 0.4062 0.8840 0.8889 0.8864 0.1282 0.1562 0.1408 0.5652 0.4483 0.5000 0.4968 0.6050
0.0608 10.0 1250 0.3169 0.5485 0.5897 0.5683 0.9289 0.4887 0.5361 0.5113 0.3333 0.3125 0.3226 0.8820 0.8722 0.8771 0.1389 0.1562 0.1471 0.3846 0.5172 0.4412 0.4598 0.5721
0.0608 11.0 1375 0.3367 0.5673 0.6227 0.5937 0.9261 0.5081 0.5794 0.5414 0.5625 0.2812 0.375 0.8798 0.8944 0.8871 0.175 0.2188 0.1944 0.35 0.4828 0.4058 0.4807 0.5966
0.0214 12.0 1500 0.3600 0.5917 0.6425 0.6161 0.9314 0.5325 0.5918 0.5605 0.4516 0.4375 0.4444 0.8684 0.9167 0.8919 0.2258 0.2188 0.2222 0.4375 0.4828 0.4590 0.5156 0.6162
0.0214 13.0 1625 0.3514 0.5606 0.6161 0.5871 0.9279 0.4882 0.5546 0.5193 0.4412 0.4688 0.4545 0.8967 0.9167 0.9066 0.1351 0.1562 0.1449 0.4815 0.4483 0.4643 0.4979 0.5906
0.0214 14.0 1750 0.3994 0.5654 0.5871 0.5761 0.9270 0.5090 0.5258 0.5172 0.3333 0.2812 0.3051 0.9034 0.8833 0.8933 0.125 0.1875 0.15 0.4571 0.5517 0.5 0.4731 0.5814
0.0214 15.0 1875 0.4133 0.5858 0.5989 0.5923 0.9292 0.5276 0.5526 0.5398 0.4737 0.2812 0.3529 0.8807 0.8611 0.8708 0.1538 0.1875 0.1690 0.4848 0.5517 0.5161 0.4897 0.5939
0.0089 16.0 2000 0.4126 0.5695 0.6108 0.5894 0.9301 0.4935 0.5505 0.5205 0.5 0.375 0.4286 0.9056 0.9056 0.9056 0.1951 0.25 0.2192 0.4815 0.4483 0.4643 0.5076 0.5932
0.0089 17.0 2125 0.4195 0.5856 0.6095 0.5973 0.9288 0.5057 0.5505 0.5271 0.6923 0.2812 0.4 0.9157 0.9056 0.9106 0.1765 0.1875 0.1818 0.4722 0.5862 0.5231 0.5085 0.5981
0.0089 18.0 2250 0.4177 0.5856 0.6227 0.6036 0.9300 0.5036 0.5711 0.5353 0.5 0.375 0.4286 0.9171 0.9222 0.9197 0.1667 0.1562 0.1613 0.5714 0.4138 0.4800 0.5050 0.6041
0.0089 19.0 2375 0.4675 0.5623 0.5897 0.5757 0.9257 0.5038 0.5402 0.5214 0.4118 0.2188 0.2857 0.9023 0.8722 0.8870 0.0943 0.1562 0.1176 0.5161 0.5517 0.5333 0.4690 0.5817
0.004 20.0 2500 0.4435 0.5604 0.6055 0.5821 0.9276 0.4878 0.5340 0.5098 0.4643 0.4062 0.4333 0.9066 0.9167 0.9116 0.15 0.1875 0.1667 0.4211 0.5517 0.4776 0.4998 0.5863
0.004 21.0 2625 0.4669 0.5516 0.5989 0.5743 0.9277 0.4822 0.5299 0.5049 0.4828 0.4375 0.4590 0.8962 0.9111 0.9036 0.1463 0.1875 0.1644 0.3514 0.4483 0.3939 0.4852 0.5790
0.004 22.0 2750 0.4732 0.5820 0.6042 0.5929 0.9285 0.5058 0.5381 0.5215 0.4643 0.4062 0.4333 0.9153 0.9 0.9076 0.2105 0.25 0.2286 0.5 0.4828 0.4912 0.5164 0.5959
0.004 23.0 2875 0.4922 0.5816 0.6016 0.5914 0.9258 0.5048 0.5402 0.5219 0.5 0.3438 0.4074 0.9091 0.8889 0.8989 0.2162 0.25 0.2319 0.5 0.5172 0.5085 0.5137 0.5938
0.0016 24.0 3000 0.4747 0.5789 0.6148 0.5963 0.9294 0.5038 0.5526 0.5270 0.4667 0.4375 0.4516 0.9056 0.9056 0.9056 0.2 0.1875 0.1935 0.4545 0.5172 0.4839 0.5123 0.5980
0.0016 25.0 3125 0.4849 0.5851 0.6121 0.5983 0.9300 0.5085 0.5526 0.5296 0.4783 0.3438 0.4 0.9011 0.9111 0.9061 0.2069 0.1875 0.1967 0.4688 0.5172 0.4918 0.5048 0.5981
0.0016 26.0 3250 0.4692 0.5821 0.6266 0.6036 0.9307 0.5009 0.5629 0.5301 0.48 0.375 0.4211 0.9176 0.9278 0.9227 0.2424 0.25 0.2462 0.4839 0.5172 0.5000 0.5240 0.6056
0.0016 27.0 3375 0.4785 0.5752 0.6108 0.5925 0.9299 0.5 0.5443 0.5212 0.4615 0.375 0.4138 0.9011 0.9111 0.9061 0.2105 0.25 0.2286 0.4839 0.5172 0.5000 0.5139 0.5949
0.001 28.0 3500 0.4873 0.5810 0.6201 0.5999 0.9322 0.5103 0.5629 0.5353 0.4815 0.4062 0.4407 0.8962 0.9111 0.9036 0.1613 0.1562 0.1587 0.4545 0.5172 0.4839 0.5044 0.6009
0.001 29.0 3625 0.4825 0.5813 0.6227 0.6013 0.9318 0.5028 0.5629 0.5311 0.52 0.4062 0.4561 0.8962 0.9111 0.9036 0.2333 0.2188 0.2258 0.4839 0.5172 0.5000 0.5233 0.6023
0.001 30.0 3750 0.4883 0.5769 0.6135 0.5946 0.9307 0.4944 0.5505 0.5210 0.52 0.4062 0.4561 0.9111 0.9111 0.9111 0.2069 0.1875 0.1967 0.4688 0.5172 0.4918 0.5154 0.5961
0.001 31.0 3875 0.4964 0.5734 0.6135 0.5927 0.9308 0.4963 0.5526 0.5229 0.5 0.4062 0.4483 0.9011 0.9111 0.9061 0.1613 0.1562 0.1587 0.4688 0.5172 0.4918 0.5056 0.5942
0.0005 32.0 4000 0.4977 0.5817 0.6201 0.6003 0.9309 0.5047 0.5588 0.5303 0.52 0.4062 0.4561 0.9066 0.9167 0.9116 0.1875 0.1875 0.1875 0.4688 0.5172 0.4918 0.5155 0.6018
0.0005 33.0 4125 0.5008 0.5810 0.6201 0.5999 0.9312 0.5047 0.5567 0.5294 0.5 0.4062 0.4483 0.9121 0.9222 0.9171 0.1765 0.1875 0.1818 0.4688 0.5172 0.4918 0.5137 0.6019
0.0005 34.0 4250 0.5028 0.5829 0.6214 0.6015 0.9310 0.5047 0.5588 0.5303 0.52 0.4062 0.4561 0.9121 0.9222 0.9171 0.1875 0.1875 0.1875 0.4688 0.5172 0.4918 0.5166 0.6031
0.0005 35.0 4375 0.5037 0.5849 0.6227 0.6032 0.9311 0.5065 0.5608 0.5323 0.52 0.4062 0.4561 0.9121 0.9222 0.9171 0.1875 0.1875 0.1875 0.4839 0.5172 0.5000 0.5186 0.6047

Framework versions

  • Transformers 4.35.2
  • Pytorch 2.1.0+cu121
  • Datasets 2.15.0
  • Tokenizers 0.15.0