<!-- 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-uncased-sst-2-32-13-smoothed
This model is a fine-tuned version of bert-base-uncased on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.6688
- Accuracy: 0.7188
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: 1e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- num_epochs: 75
- label_smoothing_factor: 0.45
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| No log | 1.0 | 2 | 0.7003 | 0.4844 |
| No log | 2.0 | 4 | 0.6998 | 0.5 |
| No log | 3.0 | 6 | 0.6991 | 0.5 |
| No log | 4.0 | 8 | 0.6980 | 0.5 |
| 0.6958 | 5.0 | 10 | 0.6967 | 0.4844 |
| 0.6958 | 6.0 | 12 | 0.6953 | 0.4844 |
| 0.6958 | 7.0 | 14 | 0.6940 | 0.5312 |
| 0.6958 | 8.0 | 16 | 0.6928 | 0.5312 |
| 0.6958 | 9.0 | 18 | 0.6915 | 0.5625 |
| 0.6939 | 10.0 | 20 | 0.6904 | 0.625 |
| 0.6939 | 11.0 | 22 | 0.6894 | 0.625 |
| 0.6939 | 12.0 | 24 | 0.6885 | 0.6094 |
| 0.6939 | 13.0 | 26 | 0.6881 | 0.5938 |
| 0.6939 | 14.0 | 28 | 0.6874 | 0.5938 |
| 0.6707 | 15.0 | 30 | 0.6870 | 0.6094 |
| 0.6707 | 16.0 | 32 | 0.6869 | 0.6562 |
| 0.6707 | 17.0 | 34 | 0.6871 | 0.6562 |
| 0.6707 | 18.0 | 36 | 0.6867 | 0.625 |
| 0.6707 | 19.0 | 38 | 0.6858 | 0.625 |
| 0.6407 | 20.0 | 40 | 0.6848 | 0.6094 |
| 0.6407 | 21.0 | 42 | 0.6840 | 0.5781 |
| 0.6407 | 22.0 | 44 | 0.6850 | 0.5781 |
| 0.6407 | 23.0 | 46 | 0.6852 | 0.5625 |
| 0.6407 | 24.0 | 48 | 0.6845 | 0.5625 |
| 0.6016 | 25.0 | 50 | 0.6853 | 0.5938 |
| 0.6016 | 26.0 | 52 | 0.6859 | 0.5938 |
| 0.6016 | 27.0 | 54 | 0.6851 | 0.5781 |
| 0.6016 | 28.0 | 56 | 0.6812 | 0.5938 |
| 0.6016 | 29.0 | 58 | 0.6793 | 0.6094 |
| 0.5645 | 30.0 | 60 | 0.6786 | 0.6094 |
| 0.5645 | 31.0 | 62 | 0.6774 | 0.625 |
| 0.5645 | 32.0 | 64 | 0.6763 | 0.6719 |
| 0.5645 | 33.0 | 66 | 0.6754 | 0.6719 |
| 0.5645 | 34.0 | 68 | 0.6751 | 0.6562 |
| 0.5434 | 35.0 | 70 | 0.6748 | 0.6719 |
| 0.5434 | 36.0 | 72 | 0.6741 | 0.7031 |
| 0.5434 | 37.0 | 74 | 0.6745 | 0.6875 |
| 0.5434 | 38.0 | 76 | 0.6752 | 0.7031 |
| 0.5434 | 39.0 | 78 | 0.6756 | 0.6719 |
| 0.5383 | 40.0 | 80 | 0.6755 | 0.6719 |
| 0.5383 | 41.0 | 82 | 0.6760 | 0.6875 |
| 0.5383 | 42.0 | 84 | 0.6778 | 0.6406 |
| 0.5383 | 43.0 | 86 | 0.6802 | 0.6406 |
| 0.5383 | 44.0 | 88 | 0.6823 | 0.6406 |
| 0.5379 | 45.0 | 90 | 0.6827 | 0.5938 |
| 0.5379 | 46.0 | 92 | 0.6815 | 0.6094 |
| 0.5379 | 47.0 | 94 | 0.6804 | 0.625 |
| 0.5379 | 48.0 | 96 | 0.6790 | 0.6406 |
| 0.5379 | 49.0 | 98 | 0.6756 | 0.6562 |
| 0.5371 | 50.0 | 100 | 0.6739 | 0.6562 |
| 0.5371 | 51.0 | 102 | 0.6726 | 0.6406 |
| 0.5371 | 52.0 | 104 | 0.6718 | 0.6719 |
| 0.5371 | 53.0 | 106 | 0.6710 | 0.7031 |
| 0.5371 | 54.0 | 108 | 0.6705 | 0.6875 |
| 0.5365 | 55.0 | 110 | 0.6701 | 0.6875 |
| 0.5365 | 56.0 | 112 | 0.6698 | 0.6875 |
| 0.5365 | 57.0 | 114 | 0.6696 | 0.7031 |
| 0.5365 | 58.0 | 116 | 0.6694 | 0.6875 |
| 0.5365 | 59.0 | 118 | 0.6693 | 0.6875 |
| 0.5369 | 60.0 | 120 | 0.6690 | 0.7031 |
| 0.5369 | 61.0 | 122 | 0.6687 | 0.7188 |
| 0.5369 | 62.0 | 124 | 0.6687 | 0.7344 |
| 0.5369 | 63.0 | 126 | 0.6688 | 0.7031 |
| 0.5369 | 64.0 | 128 | 0.6688 | 0.6875 |
| 0.5365 | 65.0 | 130 | 0.6688 | 0.7031 |
| 0.5365 | 66.0 | 132 | 0.6688 | 0.7188 |
| 0.5365 | 67.0 | 134 | 0.6688 | 0.7188 |
| 0.5365 | 68.0 | 136 | 0.6688 | 0.7188 |
| 0.5365 | 69.0 | 138 | 0.6688 | 0.7344 |
| 0.5364 | 70.0 | 140 | 0.6687 | 0.7344 |
| 0.5364 | 71.0 | 142 | 0.6687 | 0.7188 |
| 0.5364 | 72.0 | 144 | 0.6687 | 0.7188 |
| 0.5364 | 73.0 | 146 | 0.6687 | 0.7188 |
| 0.5364 | 74.0 | 148 | 0.6688 | 0.7188 |
| 0.5366 | 75.0 | 150 | 0.6688 | 0.7188 |
Framework versions
- Transformers 4.32.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.4.0
- Tokenizers 0.13.3
