<!-- 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. -->
hBERTv1_new_pretrain_48_ver2_stsb
This model is a fine-tuned version of gokuls/bert_12_layer_model_v1_complete_training_new_48 on the GLUE STSB dataset. It achieves the following results on the evaluation set:
- Loss: 2.0507
- Pearson: 0.3913
- Spearmanr: 0.3858
- Combined Score: 0.3885
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: 64
- eval_batch_size: 64
- seed: 10
- distributed_type: multi-GPU
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 15
Training results
| Training Loss | Epoch | Step | Validation Loss | Pearson | Spearmanr | Combined Score |
|---|---|---|---|---|---|---|
| 2.2439 | 1.0 | 90 | 2.2582 | 0.1118 | 0.1205 | 0.1162 |
| 1.9712 | 2.0 | 180 | 2.5514 | 0.2121 | 0.2109 | 0.2115 |
| 1.6254 | 3.0 | 270 | 2.6339 | 0.2885 | 0.2887 | 0.2886 |
| 1.2292 | 4.0 | 360 | 2.1543 | 0.3642 | 0.3666 | 0.3654 |
| 0.9444 | 5.0 | 450 | 2.6438 | 0.3529 | 0.3577 | 0.3553 |
| 0.7567 | 6.0 | 540 | 2.3755 | 0.3820 | 0.3872 | 0.3846 |
| 0.5838 | 7.0 | 630 | 2.0507 | 0.3913 | 0.3858 | 0.3885 |
| 0.5032 | 8.0 | 720 | 2.5227 | 0.4037 | 0.4071 | 0.4054 |
| 0.4112 | 9.0 | 810 | 2.1436 | 0.4072 | 0.3988 | 0.4030 |
| 0.3551 | 10.0 | 900 | 2.1501 | 0.4069 | 0.4004 | 0.4037 |
| 0.2961 | 11.0 | 990 | 2.2744 | 0.4137 | 0.4080 | 0.4109 |
| 0.256 | 12.0 | 1080 | 2.3612 | 0.4115 | 0.4038 | 0.4076 |
Framework versions
- Transformers 4.34.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.14.5
- Tokenizers 0.14.1
