Downloads · 30 days
9
14% of all-time downloads
hBERTv1_new_pretrain_48_KD_stsb
gokuls/hBERTv1_new_pretrain_48_KD_stsb
hBERTv1_new_pretrain_48_KD_stsb is a text classification model from gokuls. Use it when you need a label for a piece of text. It is set up for transformers.
should probably proofread and complete it, then remove this comment. --
All-time downloads
66
Public
Repo size
2.4 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.bin481 MB · 100%
At a glance
- Task
- Text Classification
- Library
- transformers
- Model type
- hybridbert
- Access
- Public
- Created
- Jun 15, 2023
- Updated
- Jun 15, 2023
- SHA
- 2e6c06aa
Try a prompt
Datasets
Model card
From the Hugging Face model README
<!-- 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_KD_stsb
This model is a fine-tuned version of gokuls/bert_12_layer_model_v1_complete_training_new_48_KD on the GLUE STSB dataset. It achieves the following results on the evaluation set:
- Loss: 1.9753
- Pearson: 0.4441
- Spearmanr: 0.4350
- Combined Score: 0.4395
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 |
|---|---|---|---|---|---|---|
| 2.2736 | 1.0 | 45 | 2.5807 | 0.1213 | 0.1131 | 0.1172 |
| 1.9996 | 2.0 | 90 | 2.3578 | 0.1791 | 0.1901 | 0.1846 |
| 1.7489 | 3.0 | 135 | 2.1912 | 0.2981 | 0.2958 | 0.2970 |
| 1.4124 | 4.0 | 180 | 2.8188 | 0.2915 | 0.2901 | 0.2908 |
| 1.1148 | 5.0 | 225 | 2.3077 | 0.3345 | 0.3206 | 0.3276 |
| 0.8203 | 6.0 | 270 | 2.4569 | 0.3944 | 0.3852 | 0.3898 |
| 0.6562 | 7.0 | 315 | 2.1797 | 0.4086 | 0.4082 | 0.4084 |
| 0.5537 | 8.0 | 360 | 2.2254 | 0.4198 | 0.4180 | 0.4189 |
| 0.5236 | 9.0 | 405 | 2.2477 | 0.4231 | 0.4100 | 0.4166 |
| 0.3807 | 10.0 | 450 | 2.1156 | 0.4398 | 0.4346 | 0.4372 |
| 0.3645 | 11.0 | 495 | 1.9753 | 0.4441 | 0.4350 | 0.4395 |
| 0.2975 | 12.0 | 540 | 2.3133 | 0.4634 | 0.4537 | 0.4585 |
| 0.2781 | 13.0 | 585 | 2.3479 | 0.4473 | 0.4469 | 0.4471 |
| 0.2308 | 14.0 | 630 | 2.2469 | 0.4441 | 0.4376 | 0.4408 |
| 0.2555 | 15.0 | 675 | 2.0949 | 0.4840 | 0.4772 | 0.4806 |
| 0.2003 | 16.0 | 720 | 2.0903 | 0.4814 | 0.4742 | 0.4778 |
Framework versions
- Transformers 4.30.2
- Pytorch 1.14.0a0+410ce96
- Datasets 2.12.0
- Tokenizers 0.13.3
- Task
- Text Classification
- Library
- transformers
- Type
- hybridbert
- Languages
- en
- Created
- Jun 15, 2023
- Updated
- Jun 15, 2023