scenario-non-kd-scr-ner-full_data-univner_full66
This model is a fine-tuned version of haryoaw/scenario-TCR-NER_data-univner_full on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.1213
- Precision: 0.8519
- Recall: 0.8637
- F1: 0.8577
- Accuracy: 0.9843
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: 3e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 66
- 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 | Precision | Recall | F1 | Accuracy |
---|---|---|---|---|---|---|---|
0.0089 | 0.2910 | 500 | 0.0854 | 0.8477 | 0.8498 | 0.8488 | 0.9839 |
0.0113 | 0.5821 | 1000 | 0.0804 | 0.8488 | 0.8660 | 0.8573 | 0.9846 |
0.0106 | 0.8731 | 1500 | 0.0797 | 0.8506 | 0.8665 | 0.8585 | 0.9845 |
0.0094 | 1.1641 | 2000 | 0.0869 | 0.8584 | 0.8622 | 0.8603 | 0.9847 |
0.0081 | 1.4552 | 2500 | 0.0932 | 0.8417 | 0.8664 | 0.8539 | 0.9839 |
0.0093 | 1.7462 | 3000 | 0.0842 | 0.8416 | 0.8673 | 0.8543 | 0.9842 |
0.0075 | 2.0373 | 3500 | 0.0912 | 0.8355 | 0.8691 | 0.8520 | 0.9838 |
0.0061 | 2.3283 | 4000 | 0.0825 | 0.8491 | 0.8518 | 0.8505 | 0.9842 |
0.007 | 2.6193 | 4500 | 0.0914 | 0.8494 | 0.8569 | 0.8531 | 0.9845 |
0.0072 | 2.9104 | 5000 | 0.0940 | 0.8481 | 0.8619 | 0.8550 | 0.9842 |
0.005 | 3.2014 | 5500 | 0.0922 | 0.8509 | 0.8588 | 0.8548 | 0.9845 |
0.0051 | 3.4924 | 6000 | 0.0985 | 0.8414 | 0.8600 | 0.8506 | 0.9839 |
0.0055 | 3.7835 | 6500 | 0.0878 | 0.8466 | 0.8517 | 0.8491 | 0.9843 |
0.0054 | 4.0745 | 7000 | 0.1042 | 0.8587 | 0.8674 | 0.8630 | 0.9850 |
0.0045 | 4.3655 | 7500 | 0.0981 | 0.8534 | 0.8639 | 0.8586 | 0.9841 |
0.0054 | 4.6566 | 8000 | 0.1017 | 0.8557 | 0.8665 | 0.8611 | 0.9849 |
0.0046 | 4.9476 | 8500 | 0.1031 | 0.8396 | 0.8634 | 0.8513 | 0.9834 |
0.0041 | 5.2386 | 9000 | 0.1015 | 0.8475 | 0.8603 | 0.8539 | 0.9842 |
0.0039 | 5.5297 | 9500 | 0.1053 | 0.8500 | 0.8628 | 0.8564 | 0.9846 |
0.0037 | 5.8207 | 10000 | 0.1113 | 0.8478 | 0.8654 | 0.8565 | 0.9845 |
0.0045 | 6.1118 | 10500 | 0.1221 | 0.8432 | 0.8714 | 0.8571 | 0.9841 |
0.0041 | 6.4028 | 11000 | 0.1001 | 0.8518 | 0.8639 | 0.8578 | 0.9844 |
0.0031 | 6.6938 | 11500 | 0.1068 | 0.8564 | 0.8632 | 0.8598 | 0.9846 |
0.0037 | 6.9849 | 12000 | 0.1108 | 0.8450 | 0.8613 | 0.8531 | 0.9841 |
0.0027 | 7.2759 | 12500 | 0.1184 | 0.8463 | 0.8647 | 0.8554 | 0.9838 |
0.0031 | 7.5669 | 13000 | 0.1205 | 0.8460 | 0.8611 | 0.8535 | 0.9837 |
0.0031 | 7.8580 | 13500 | 0.1132 | 0.8486 | 0.8719 | 0.8601 | 0.9845 |
0.003 | 8.1490 | 14000 | 0.1104 | 0.8440 | 0.8600 | 0.8519 | 0.9840 |
0.0021 | 8.4400 | 14500 | 0.1214 | 0.8529 | 0.8469 | 0.8499 | 0.9838 |
0.0029 | 8.7311 | 15000 | 0.1154 | 0.8408 | 0.8559 | 0.8483 | 0.9837 |
0.0028 | 9.0221 | 15500 | 0.1117 | 0.8460 | 0.8686 | 0.8571 | 0.9845 |
0.0021 | 9.3132 | 16000 | 0.1253 | 0.8347 | 0.8694 | 0.8517 | 0.9836 |
0.0018 | 9.6042 | 16500 | 0.1239 | 0.8487 | 0.8663 | 0.8574 | 0.9844 |
0.0021 | 9.8952 | 17000 | 0.1213 | 0.8519 | 0.8637 | 0.8577 | 0.9843 |
Framework versions
- Transformers 4.44.2
- Pytorch 2.1.1+cu121
- Datasets 2.14.5
- Tokenizers 0.19.1
- Downloads last month
- 14
Inference API (serverless) is not available, repository is disabled.
Model tree for haryoaw/scenario-non-kd-scr-ner-full_data-univner_full66
Base model
FacebookAI/xlm-roberta-base
Finetuned
this model