<!-- 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. -->
small-mlm-glue-qnli-custom-tokenizer
This model is a fine-tuned version of google/bert_uncased_L-4_H-512_A-8 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 5.5974
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: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- num_epochs: 200
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 7.2117 | 0.4 | 500 | 6.7997 |
| 6.5734 | 0.8 | 1000 | 6.6026 |
| 6.4063 | 1.2 | 1500 | 6.5514 |
| 6.2622 | 1.6 | 2000 | 6.4092 |
| 6.2126 | 2.0 | 2500 | 6.3653 |
| 6.1191 | 2.4 | 3000 | 6.3054 |
| 6.0669 | 2.8 | 3500 | 6.2685 |
| 5.9877 | 3.2 | 4000 | 6.2077 |
| 5.8901 | 3.6 | 4500 | 6.1328 |
| 5.8306 | 4.0 | 5000 | 6.1574 |
| 5.8053 | 4.4 | 5500 | 6.0958 |
| 5.7117 | 4.8 | 6000 | 6.0377 |
| 5.7372 | 5.2 | 6500 | 6.0045 |
| 5.6595 | 5.6 | 7000 | 5.9655 |
| 5.6579 | 6.0 | 7500 | 5.9410 |
| 5.6323 | 6.4 | 8000 | 5.9121 |
| 5.5978 | 6.8 | 8500 | 5.8435 |
| 5.5634 | 7.2 | 9000 | 5.9205 |
| 5.4642 | 7.6 | 9500 | 5.8433 |
| 5.4851 | 8.0 | 10000 | 5.8122 |
| 5.4272 | 8.4 | 10500 | 5.8350 |
| 5.357 | 8.8 | 11000 | 5.7860 |
| 5.3638 | 9.2 | 11500 | 5.7262 |
| 5.3088 | 9.6 | 12000 | 5.7529 |
| 5.3052 | 10.0 | 12500 | 5.7783 |
| 5.2628 | 10.4 | 13000 | 5.7124 |
| 5.2923 | 10.8 | 13500 | 5.7053 |
| 5.1727 | 11.2 | 14000 | 5.7031 |
| 5.1474 | 11.6 | 14500 | 5.6445 |
| 5.145 | 12.0 | 15000 | 5.6299 |
| 5.102 | 12.4 | 15500 | 5.6858 |
| 5.0612 | 12.8 | 16000 | 5.6089 |
| 5.0928 | 13.2 | 16500 | 5.6404 |
| 4.9953 | 13.6 | 17000 | 5.5769 |
| 5.0163 | 14.0 | 17500 | 5.5935 |
| 4.9591 | 14.4 | 18000 | 5.5862 |
| 5.0046 | 14.8 | 18500 | 5.5974 |
Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.13.0+cu116
- Datasets 2.8.1.dev0
- Tokenizers 0.13.2
