<!-- 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. -->
hBERTv2_new_pretrain_48_KD_w_init_qnli
This model is a fine-tuned version of gokuls/bert_12_layer_model_v2_complete_training_new_48_KD_wt_init on the GLUE QNLI dataset. It achieves the following results on the evaluation set:
- Loss: 0.6306
- Accuracy: 0.6418
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: 4e-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 |
---|---|---|---|---|
0.6733 | 1.0 | 819 | 0.6643 | 0.5913 |
0.641 | 2.0 | 1638 | 0.6306 | 0.6418 |
0.5952 | 3.0 | 2457 | 0.6488 | 0.6377 |
0.5439 | 4.0 | 3276 | 0.6661 | 0.6302 |
0.4907 | 5.0 | 4095 | 0.6937 | 0.6253 |
0.4364 | 6.0 | 4914 | 0.7381 | 0.6297 |
0.3825 | 7.0 | 5733 | 0.8475 | 0.6240 |
Framework versions
- Transformers 4.30.2
- Pytorch 1.14.0a0+410ce96
- Datasets 2.13.0
- Tokenizers 0.13.3