<!-- 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_emb_com_stsb
This model is a fine-tuned version of gokuls/bert_12_layer_model_v2_complete_training_new_emb_compress_48 on the GLUE STSB dataset. It achieves the following results on the evaluation set:
- Loss: 2.0889
- Pearson: 0.3123
- Spearmanr: 0.3073
- Combined Score: 0.3098
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 | Pearson | Spearmanr | Combined Score |
|---|---|---|---|---|---|---|
| 2.398 | 1.0 | 45 | 3.0621 | 0.0972 | 0.1007 | 0.0990 |
| 2.0392 | 2.0 | 90 | 2.3674 | 0.1058 | 0.1011 | 0.1034 |
| 1.967 | 3.0 | 135 | 2.2296 | 0.1449 | 0.1432 | 0.1441 |
| 1.8176 | 4.0 | 180 | 2.6036 | 0.2055 | 0.2169 | 0.2112 |
| 1.6744 | 5.0 | 225 | 2.2119 | 0.2516 | 0.2534 | 0.2525 |
| 1.4727 | 6.0 | 270 | 2.0889 | 0.3123 | 0.3073 | 0.3098 |
| 1.1852 | 7.0 | 315 | 2.6372 | 0.3609 | 0.3543 | 0.3576 |
| 0.9895 | 8.0 | 360 | 2.5881 | 0.3312 | 0.3322 | 0.3317 |
| 0.8254 | 9.0 | 405 | 2.1746 | 0.3991 | 0.3974 | 0.3983 |
| 0.6759 | 10.0 | 450 | 2.7671 | 0.3693 | 0.3663 | 0.3678 |
| 0.558 | 11.0 | 495 | 2.5954 | 0.3967 | 0.3942 | 0.3955 |
Framework versions
- Transformers 4.30.2
- Pytorch 1.14.0a0+410ce96
- Datasets 2.12.0
- Tokenizers 0.13.3
