README.md

metadata

language:
  - en
tags:
  - generated_from_trainer
datasets:
  - glue
metrics:
  - spearmanr
model-index:
  - name: hBERTv1_stsb
    results:
      - task:
          name: Text Classification
          type: text-classification
        dataset:
          name: GLUE STSB
          type: glue
          config: stsb
          split: validation
          args: stsb
        metrics:
          - name: Spearmanr
            type: spearmanr
            value: 0.7155715863961268

hBERTv1_stsb

This model is a fine-tuned version of gokuls/bert_12_layer_model_v1 on the GLUE STSB dataset. It achieves the following results on the evaluation set:

• Loss: 1.1154
• Pearson: 0.7159
• Spearmanr: 0.7156
• Combined Score: 0.7157

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: 256
• eval_batch_size: 256
• 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	Pearson	Spearmanr	Combined Score
4.0796	1.0	23	2.3017	0.0761	0.0547	0.0654
2.0746	2.0	46	2.6181	0.0850	0.0772	0.0811
1.9142	3.0	69	2.2963	0.1878	0.1852	0.1865
1.6883	4.0	92	2.1866	0.4740	0.4777	0.4759
1.1166	5.0	115	1.9367	0.6319	0.6450	0.6384
0.7598	6.0	138	1.4188	0.6801	0.6888	0.6845
0.5453	7.0	161	1.2720	0.6988	0.7001	0.6994
0.3705	8.0	184	1.1154	0.7159	0.7156	0.7157
0.2976	9.0	207	1.6889	0.6754	0.6807	0.6780
0.2272	10.0	230	1.3627	0.6929	0.6899	0.6914
0.1966	11.0	253	1.1278	0.7195	0.7167	0.7181
0.1708	12.0	276	1.3476	0.7171	0.7165	0.7168
0.1529	13.0	299	1.2614	0.6982	0.6942	0.6962

Framework versions

• Transformers 4.26.1
• Pytorch 1.14.0a0+410ce96
• Datasets 2.10.1
• Tokenizers 0.13.2