<!-- 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
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: 0.4287
- Accuracy: 0.8007
- F1: 0.7301
- Combined Score: 0.7654
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
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Combined Score |
---|---|---|---|---|---|---|
0.5237 | 1.0 | 2843 | 0.5087 | 0.7472 | 0.6697 | 0.7085 |
0.4614 | 2.0 | 5686 | 0.4697 | 0.7754 | 0.6746 | 0.7250 |
0.4287 | 3.0 | 8529 | 0.4508 | 0.7853 | 0.6893 | 0.7373 |
0.4089 | 4.0 | 11372 | 0.4493 | 0.7925 | 0.7151 | 0.7538 |
0.3904 | 5.0 | 14215 | 0.4361 | 0.7984 | 0.7222 | 0.7603 |
0.3752 | 6.0 | 17058 | 0.4332 | 0.8023 | 0.7215 | 0.7619 |
0.3592 | 7.0 | 19901 | 0.4287 | 0.8007 | 0.7301 | 0.7654 |
0.3458 | 8.0 | 22744 | 0.4337 | 0.8005 | 0.7324 | 0.7664 |
0.3326 | 9.0 | 25587 | 0.4340 | 0.8006 | 0.7362 | 0.7684 |
0.3201 | 10.0 | 28430 | 0.4464 | 0.8028 | 0.7417 | 0.7722 |
0.3092 | 11.0 | 31273 | 0.4615 | 0.8037 | 0.7196 | 0.7617 |
0.2984 | 12.0 | 34116 | 0.4763 | 0.8047 | 0.7326 | 0.7687 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.8.0
- Tokenizers 0.13.2