bert_uncased_L-12_H-128_A-2-mlm-multi-emails-hq
This model is a fine-tuned version of google/bert_uncased_L-12_H-128_A-2 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 2.4235
- Accuracy: 0.5780
Model description
This is a 40 MB version of BERT that does surprisingly well!
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: 0.0003
- train_batch_size: 8
- eval_batch_size: 2
- seed: 42
- distributed_type: multi-GPU
- gradient_accumulation_steps: 16
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.05
- num_epochs: 8.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 2.9421 | 0.99 | 141 | 2.7769 | 0.5330 |
| 2.772 | 1.99 | 282 | 2.6669 | 0.5487 |
| 2.6997 | 2.99 | 423 | 2.5486 | 0.5621 |
| 2.6281 | 3.99 | 564 | 2.4865 | 0.5704 |
| 2.5626 | 4.99 | 705 | 2.4385 | 0.5766 |
| 2.5504 | 5.99 | 846 | 2.4421 | 0.5772 |
| 2.5434 | 6.99 | 987 | 2.4094 | 0.5818 |
| 2.5174 | 7.99 | 1128 | 2.4235 | 0.5780 |
Framework versions
- Transformers 4.27.0.dev0
- Pytorch 2.0.0.dev20230129+cu118
- Datasets 2.8.0
- Tokenizers 0.13.1