mobilebert_add_GLUE_Experiment_logit_kd_sst2_128
This model is a fine-tuned version of google/mobilebert-uncased on the GLUE SST2 dataset. It achieves the following results on the evaluation set:
- Loss: 0.7282
- Accuracy: 0.8073
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: 128
- eval_batch_size: 128
- seed: 10
- distributed_type: multi-GPU
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 50
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.5487 | 1.0 | 527 | 1.3929 | 0.5780 |
| 1.3629 | 2.0 | 1054 | 1.4979 | 0.5505 |
| 1.1397 | 3.0 | 1581 | 1.3927 | 0.6755 |
| 0.5649 | 4.0 | 2108 | 0.7289 | 0.8073 |
| 0.4112 | 5.0 | 2635 | 0.7282 | 0.8073 |
| 0.3462 | 6.0 | 3162 | 0.7654 | 0.8050 |
| 0.3069 | 7.0 | 3689 | 0.8303 | 0.7970 |
| 0.2833 | 8.0 | 4216 | 0.8806 | 0.7924 |
| 0.2662 | 9.0 | 4743 | 0.9297 | 0.7959 |
| 0.2521 | 10.0 | 5270 | 1.0979 | 0.7718 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.9.0
- Tokenizers 0.13.2