<!-- 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. -->
mobilebert_sa_GLUE_Experiment_qqp_128
This model is a fine-tuned version of google/mobilebert-uncased on the GLUE QQP dataset. It achieves the following results on the evaluation set:
- Loss: 0.4700
- Accuracy: 0.7784
- F1: 0.6886
- Combined Score: 0.7335
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 | F1 | Combined Score |
|---|---|---|---|---|---|---|
| 0.5294 | 1.0 | 2843 | 0.5076 | 0.7512 | 0.6636 | 0.7074 |
| 0.4791 | 2.0 | 5686 | 0.4889 | 0.7613 | 0.6369 | 0.6991 |
| 0.4622 | 3.0 | 8529 | 0.4821 | 0.7657 | 0.6475 | 0.7066 |
| 0.4463 | 4.0 | 11372 | 0.4831 | 0.7694 | 0.6730 | 0.7212 |
| 0.4288 | 5.0 | 14215 | 0.4724 | 0.7752 | 0.6784 | 0.7268 |
| 0.4129 | 6.0 | 17058 | 0.4806 | 0.7749 | 0.6893 | 0.7321 |
| 0.3969 | 7.0 | 19901 | 0.4700 | 0.7784 | 0.6886 | 0.7335 |
| 0.3813 | 8.0 | 22744 | 0.4802 | 0.7790 | 0.6962 | 0.7376 |
| 0.3664 | 9.0 | 25587 | 0.4765 | 0.7805 | 0.6952 | 0.7378 |
| 0.352 | 10.0 | 28430 | 0.4965 | 0.7768 | 0.7086 | 0.7427 |
| 0.3381 | 11.0 | 31273 | 0.4895 | 0.7845 | 0.6960 | 0.7403 |
| 0.3258 | 12.0 | 34116 | 0.5092 | 0.7844 | 0.7043 | 0.7444 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.8.0
- Tokenizers 0.13.2
