File size: 14,783 Bytes
70f4ff8 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 |
---
license: apache-2.0
base_model: bert-base-cased
tags:
- generated_from_trainer
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: bert-base-cased-finetuned-ner-cadec-no-iob
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bert-base-cased-finetuned-ner-cadec-no-iob
This model is a fine-tuned version of [bert-base-cased](https://huggingface.co/bert-base-cased) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4487
- Precision: 0.6037
- Recall: 0.6491
- F1: 0.6256
- Accuracy: 0.9313
- Adr Precision: 0.5441
- Adr Recall: 0.6103
- Adr F1: 0.5753
- Disease Precision: 0.5
- Disease Recall: 0.375
- Disease F1: 0.4286
- Drug Precision: 0.8649
- Drug Recall: 0.8889
- Drug F1: 0.8767
- Finding Precision: 0.2903
- Finding Recall: 0.2812
- Finding F1: 0.2857
- Symptom Precision: 0.4839
- Symptom Recall: 0.5172
- Symptom F1: 0.5000
- Macro Avg F1: 0.5333
- Weighted Avg F1: 0.6256
## 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 | 0.2340 | 0.5044 | 0.6003 | 0.5482 | 0.9191 | 0.4397 | 0.5938 | 0.5053 | 0.3529 | 0.375 | 0.3636 | 0.7574 | 0.85 | 0.8010 | 0.1818 | 0.0625 | 0.0930 | 0.0 | 0.0 | 0.0 | 0.3526 | 0.5328 |
| No log | 2.0 | 250 | 0.2068 | 0.5546 | 0.6227 | 0.5867 | 0.9253 | 0.4770 | 0.6 | 0.5315 | 0.55 | 0.3438 | 0.4231 | 0.8256 | 0.8944 | 0.8587 | 0.3158 | 0.1875 | 0.2353 | 0.4286 | 0.1034 | 0.1667 | 0.4430 | 0.5782 |
| No log | 3.0 | 375 | 0.2031 | 0.5633 | 0.6161 | 0.5885 | 0.9281 | 0.5150 | 0.5670 | 0.5397 | 0.5 | 0.4062 | 0.4483 | 0.8093 | 0.8722 | 0.8396 | 0.2 | 0.2188 | 0.2090 | 0.375 | 0.5172 | 0.4348 | 0.4943 | 0.5891 |
| 0.209 | 4.0 | 500 | 0.2331 | 0.5483 | 0.6069 | 0.5761 | 0.9273 | 0.5009 | 0.5897 | 0.5417 | 0.0 | 0.0 | 0.0 | 0.8404 | 0.8778 | 0.8587 | 0.14 | 0.2188 | 0.1707 | 0.5 | 0.3103 | 0.3830 | 0.3908 | 0.5724 |
| 0.209 | 5.0 | 625 | 0.2376 | 0.5878 | 0.6491 | 0.6169 | 0.9324 | 0.5129 | 0.6165 | 0.5599 | 0.5312 | 0.5312 | 0.5312 | 0.8703 | 0.8944 | 0.8822 | 0.1429 | 0.0625 | 0.0870 | 0.5652 | 0.4483 | 0.5000 | 0.5121 | 0.6130 |
| 0.209 | 6.0 | 750 | 0.2523 | 0.5646 | 0.6346 | 0.5975 | 0.9258 | 0.5114 | 0.6021 | 0.5530 | 0.4 | 0.375 | 0.3871 | 0.8649 | 0.8889 | 0.8767 | 0.0857 | 0.0938 | 0.0896 | 0.4516 | 0.4828 | 0.4667 | 0.4746 | 0.6000 |
| 0.209 | 7.0 | 875 | 0.2753 | 0.5748 | 0.6438 | 0.6073 | 0.9249 | 0.5209 | 0.6165 | 0.5647 | 0.4762 | 0.3125 | 0.3774 | 0.8670 | 0.9056 | 0.8859 | 0.1458 | 0.2188 | 0.1750 | 0.5 | 0.3103 | 0.3830 | 0.4772 | 0.6096 |
| 0.0561 | 8.0 | 1000 | 0.2769 | 0.5868 | 0.6557 | 0.6193 | 0.9284 | 0.5288 | 0.6247 | 0.5728 | 0.6 | 0.375 | 0.4615 | 0.8703 | 0.8944 | 0.8822 | 0.2424 | 0.25 | 0.2462 | 0.3611 | 0.4483 | 0.4000 | 0.5125 | 0.6212 |
| 0.0561 | 9.0 | 1125 | 0.3161 | 0.5719 | 0.6240 | 0.5968 | 0.9281 | 0.5091 | 0.5794 | 0.5419 | 0.5263 | 0.3125 | 0.3922 | 0.8757 | 0.9 | 0.8877 | 0.1739 | 0.25 | 0.2051 | 0.48 | 0.4138 | 0.4444 | 0.4943 | 0.5998 |
| 0.0561 | 10.0 | 1250 | 0.3101 | 0.5867 | 0.6385 | 0.6115 | 0.9297 | 0.5343 | 0.5938 | 0.5625 | 0.4839 | 0.4688 | 0.4762 | 0.8791 | 0.8889 | 0.8840 | 0.1818 | 0.25 | 0.2105 | 0.4483 | 0.4483 | 0.4483 | 0.5163 | 0.6160 |
| 0.0561 | 11.0 | 1375 | 0.3321 | 0.5862 | 0.6412 | 0.6125 | 0.9295 | 0.5245 | 0.5959 | 0.5579 | 0.6 | 0.4688 | 0.5263 | 0.8556 | 0.8889 | 0.8719 | 0.2286 | 0.25 | 0.2388 | 0.4516 | 0.4828 | 0.4667 | 0.5323 | 0.6142 |
| 0.0206 | 12.0 | 1500 | 0.3459 | 0.5923 | 0.6517 | 0.6206 | 0.9303 | 0.5323 | 0.6124 | 0.5695 | 0.5517 | 0.5 | 0.5246 | 0.875 | 0.8944 | 0.8846 | 0.2581 | 0.25 | 0.2540 | 0.375 | 0.4138 | 0.3934 | 0.5252 | 0.6224 |
| 0.0206 | 13.0 | 1625 | 0.3489 | 0.5866 | 0.6214 | 0.6035 | 0.9270 | 0.5327 | 0.5876 | 0.5588 | 0.4667 | 0.4375 | 0.4516 | 0.8370 | 0.8556 | 0.8462 | 0.24 | 0.1875 | 0.2105 | 0.4138 | 0.4138 | 0.4138 | 0.4962 | 0.6023 |
| 0.0206 | 14.0 | 1750 | 0.3762 | 0.5709 | 0.6214 | 0.5951 | 0.9270 | 0.5047 | 0.5588 | 0.5303 | 0.5 | 0.4375 | 0.4667 | 0.8811 | 0.9056 | 0.8932 | 0.2143 | 0.2812 | 0.2432 | 0.4242 | 0.4828 | 0.4516 | 0.5170 | 0.5987 |
| 0.0206 | 15.0 | 1875 | 0.3729 | 0.5806 | 0.6412 | 0.6094 | 0.9280 | 0.5149 | 0.6041 | 0.5560 | 0.5652 | 0.4062 | 0.4727 | 0.8503 | 0.8833 | 0.8665 | 0.3 | 0.2812 | 0.2903 | 0.4286 | 0.4138 | 0.4211 | 0.5213 | 0.6098 |
| 0.0093 | 16.0 | 2000 | 0.3980 | 0.5748 | 0.6385 | 0.6050 | 0.9265 | 0.5229 | 0.6124 | 0.5641 | 0.4762 | 0.3125 | 0.3774 | 0.8525 | 0.8667 | 0.8595 | 0.2326 | 0.3125 | 0.2667 | 0.4074 | 0.3793 | 0.3929 | 0.4921 | 0.6073 |
| 0.0093 | 17.0 | 2125 | 0.3885 | 0.5951 | 0.6359 | 0.6148 | 0.9285 | 0.5343 | 0.5938 | 0.5625 | 0.6087 | 0.4375 | 0.5091 | 0.8587 | 0.8778 | 0.8681 | 0.25 | 0.25 | 0.25 | 0.4375 | 0.4828 | 0.4590 | 0.5297 | 0.6157 |
| 0.0093 | 18.0 | 2250 | 0.4024 | 0.6015 | 0.6491 | 0.6244 | 0.9310 | 0.5368 | 0.6021 | 0.5675 | 0.5 | 0.4375 | 0.4667 | 0.8811 | 0.9056 | 0.8932 | 0.2857 | 0.25 | 0.2667 | 0.4545 | 0.5172 | 0.4839 | 0.5356 | 0.6247 |
| 0.0093 | 19.0 | 2375 | 0.4019 | 0.6025 | 0.6478 | 0.6243 | 0.9302 | 0.5399 | 0.6 | 0.5684 | 0.5714 | 0.5 | 0.5333 | 0.8703 | 0.8944 | 0.8822 | 0.2667 | 0.25 | 0.2581 | 0.4545 | 0.5172 | 0.4839 | 0.5452 | 0.6251 |
| 0.0053 | 20.0 | 2500 | 0.4061 | 0.5847 | 0.6332 | 0.6080 | 0.9291 | 0.5268 | 0.5876 | 0.5556 | 0.5652 | 0.4062 | 0.4727 | 0.8595 | 0.8833 | 0.8712 | 0.2286 | 0.25 | 0.2388 | 0.4054 | 0.5172 | 0.4545 | 0.5186 | 0.6098 |
| 0.0053 | 21.0 | 2625 | 0.4219 | 0.5903 | 0.6425 | 0.6153 | 0.9288 | 0.5213 | 0.6062 | 0.5605 | 0.55 | 0.3438 | 0.4231 | 0.8587 | 0.8778 | 0.8681 | 0.3103 | 0.2812 | 0.2951 | 0.5357 | 0.5172 | 0.5263 | 0.5346 | 0.6153 |
| 0.0053 | 22.0 | 2750 | 0.4190 | 0.6024 | 0.6557 | 0.6279 | 0.9309 | 0.5420 | 0.6247 | 0.5805 | 0.5185 | 0.4375 | 0.4746 | 0.8548 | 0.8833 | 0.8689 | 0.32 | 0.25 | 0.2807 | 0.4643 | 0.4483 | 0.4561 | 0.5321 | 0.6271 |
| 0.0053 | 23.0 | 2875 | 0.4272 | 0.5870 | 0.6412 | 0.6129 | 0.9287 | 0.5192 | 0.5856 | 0.5504 | 0.6154 | 0.5 | 0.5517 | 0.8610 | 0.8944 | 0.8774 | 0.2564 | 0.3125 | 0.2817 | 0.5172 | 0.5172 | 0.5172 | 0.5557 | 0.6155 |
| 0.0034 | 24.0 | 3000 | 0.4206 | 0.5887 | 0.6438 | 0.6150 | 0.9308 | 0.5160 | 0.6 | 0.5548 | 0.5769 | 0.4688 | 0.5172 | 0.8602 | 0.8889 | 0.8743 | 0.2963 | 0.25 | 0.2712 | 0.5385 | 0.4828 | 0.5091 | 0.5453 | 0.6154 |
| 0.0034 | 25.0 | 3125 | 0.4260 | 0.6037 | 0.6491 | 0.6256 | 0.9309 | 0.5365 | 0.6062 | 0.5692 | 0.52 | 0.4062 | 0.4561 | 0.8859 | 0.9056 | 0.8956 | 0.2692 | 0.2188 | 0.2414 | 0.4688 | 0.5172 | 0.4918 | 0.5308 | 0.6251 |
| 0.0034 | 26.0 | 3250 | 0.4341 | 0.5995 | 0.6478 | 0.6227 | 0.9310 | 0.5307 | 0.6062 | 0.5659 | 0.5417 | 0.4062 | 0.4643 | 0.8710 | 0.9 | 0.8852 | 0.2857 | 0.25 | 0.2667 | 0.5185 | 0.4828 | 0.5 | 0.5364 | 0.6223 |
| 0.0034 | 27.0 | 3375 | 0.4476 | 0.6010 | 0.6438 | 0.6217 | 0.9300 | 0.5314 | 0.5938 | 0.5609 | 0.56 | 0.4375 | 0.4912 | 0.8710 | 0.9 | 0.8852 | 0.3 | 0.2812 | 0.2903 | 0.5172 | 0.5172 | 0.5172 | 0.5490 | 0.6219 |
| 0.0025 | 28.0 | 3500 | 0.4281 | 0.6010 | 0.6478 | 0.6235 | 0.9299 | 0.5328 | 0.6021 | 0.5653 | 0.56 | 0.4375 | 0.4912 | 0.8663 | 0.9 | 0.8828 | 0.2667 | 0.25 | 0.2581 | 0.5556 | 0.5172 | 0.5357 | 0.5466 | 0.6235 |
| 0.0025 | 29.0 | 3625 | 0.4339 | 0.5988 | 0.6438 | 0.6205 | 0.9299 | 0.5378 | 0.6021 | 0.5681 | 0.52 | 0.4062 | 0.4561 | 0.8595 | 0.8833 | 0.8712 | 0.2903 | 0.2812 | 0.2857 | 0.4839 | 0.5172 | 0.5000 | 0.5362 | 0.6208 |
| 0.0025 | 30.0 | 3750 | 0.4408 | 0.6105 | 0.6596 | 0.6341 | 0.9311 | 0.5404 | 0.6206 | 0.5777 | 0.5909 | 0.4062 | 0.4815 | 0.8663 | 0.9 | 0.8828 | 0.36 | 0.2812 | 0.3158 | 0.5357 | 0.5172 | 0.5263 | 0.5568 | 0.6331 |
| 0.0025 | 31.0 | 3875 | 0.4450 | 0.6079 | 0.6504 | 0.6284 | 0.9309 | 0.5410 | 0.6124 | 0.5745 | 0.5417 | 0.4062 | 0.4643 | 0.8656 | 0.8944 | 0.8798 | 0.2917 | 0.2188 | 0.25 | 0.5357 | 0.5172 | 0.5263 | 0.5390 | 0.6268 |
| 0.0016 | 32.0 | 4000 | 0.4435 | 0.5988 | 0.6359 | 0.6168 | 0.9305 | 0.5345 | 0.5918 | 0.5616 | 0.52 | 0.4062 | 0.4561 | 0.8641 | 0.8833 | 0.8736 | 0.2857 | 0.25 | 0.2667 | 0.4839 | 0.5172 | 0.5000 | 0.5316 | 0.6165 |
| 0.0016 | 33.0 | 4125 | 0.4448 | 0.6017 | 0.6438 | 0.6221 | 0.9308 | 0.5369 | 0.6 | 0.5667 | 0.5417 | 0.4062 | 0.4643 | 0.8696 | 0.8889 | 0.8791 | 0.3103 | 0.2812 | 0.2951 | 0.4688 | 0.5172 | 0.4918 | 0.5394 | 0.6222 |
| 0.0016 | 34.0 | 4250 | 0.4459 | 0.6030 | 0.6451 | 0.6233 | 0.9304 | 0.5436 | 0.6041 | 0.5723 | 0.5 | 0.375 | 0.4286 | 0.8649 | 0.8889 | 0.8767 | 0.2812 | 0.2812 | 0.2812 | 0.4839 | 0.5172 | 0.5000 | 0.5318 | 0.6234 |
| 0.0016 | 35.0 | 4375 | 0.4487 | 0.6037 | 0.6491 | 0.6256 | 0.9313 | 0.5441 | 0.6103 | 0.5753 | 0.5 | 0.375 | 0.4286 | 0.8649 | 0.8889 | 0.8767 | 0.2903 | 0.2812 | 0.2857 | 0.4839 | 0.5172 | 0.5000 | 0.5333 | 0.6256 |
### Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0
|