<!-- 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_sm_gen1_large
This model is a fine-tuned version of bert-base-uncased on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.1722
- Accuracy: 0.8151
- Precision: 0.5531
- Recall: 0.4006
- F1: 0.4647
- D-index: 1.5910
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: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 96000
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 | D-index |
|---|---|---|---|---|---|---|---|---|
| 0.4737 | 1.0 | 1500 | 0.4508 | 0.7995 | 0.0 | 0.0 | 0.0 | 1.4325 |
| 0.4244 | 2.0 | 3000 | 0.4167 | 0.8041 | 0.6514 | 0.0474 | 0.0884 | 1.4558 |
| 0.3988 | 3.0 | 4500 | 0.4196 | 0.823 | 0.6737 | 0.2259 | 0.3383 | 1.5437 |
| 0.3789 | 4.0 | 6000 | 0.4097 | 0.8283 | 0.6564 | 0.3003 | 0.4121 | 1.5758 |
| 0.3565 | 5.0 | 7500 | 0.4138 | 0.8302 | 0.6614 | 0.3120 | 0.4240 | 1.5821 |
| 0.3248 | 6.0 | 9000 | 0.4094 | 0.826 | 0.5914 | 0.4251 | 0.4947 | 1.6135 |
| 0.2971 | 7.0 | 10500 | 0.4413 | 0.8291 | 0.6156 | 0.3910 | 0.4782 | 1.6066 |
| 0.2596 | 8.0 | 12000 | 0.5242 | 0.828 | 0.6206 | 0.3640 | 0.4588 | 1.5963 |
| 0.2266 | 9.0 | 13500 | 0.5538 | 0.8266 | 0.6133 | 0.3636 | 0.4565 | 1.5943 |
| 0.185 | 10.0 | 15000 | 0.7189 | 0.8206 | 0.6390 | 0.2400 | 0.3490 | 1.5452 |
| 0.1577 | 11.0 | 16500 | 0.6479 | 0.8176 | 0.5522 | 0.4734 | 0.5097 | 1.6178 |
| 0.126 | 12.0 | 18000 | 0.7646 | 0.8204 | 0.5915 | 0.3349 | 0.4276 | 1.5766 |
| 0.1149 | 13.0 | 19500 | 0.8697 | 0.8183 | 0.5874 | 0.3132 | 0.4086 | 1.5666 |
| 0.1041 | 14.0 | 21000 | 0.8395 | 0.8134 | 0.5485 | 0.3881 | 0.4546 | 1.5847 |
| 0.0879 | 15.0 | 22500 | 1.0069 | 0.81 | 0.5376 | 0.3686 | 0.4373 | 1.5737 |
| 0.0895 | 16.0 | 24000 | 1.0749 | 0.8122 | 0.5379 | 0.4426 | 0.4856 | 1.6007 |
| 0.0784 | 17.0 | 25500 | 1.1892 | 0.8177 | 0.6040 | 0.2621 | 0.3655 | 1.5488 |
| 0.0763 | 18.0 | 27000 | 1.2229 | 0.8169 | 0.6095 | 0.2396 | 0.3440 | 1.5401 |
| 0.0675 | 19.0 | 28500 | 1.2429 | 0.8196 | 0.6038 | 0.2891 | 0.3910 | 1.5603 |
| 0.0645 | 20.0 | 30000 | 1.1722 | 0.8151 | 0.5531 | 0.4006 | 0.4647 | 1.5910 |
Framework versions
- Transformers 4.28.0
- Pytorch 2.0.1+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3
