hBERTv1_new_pretrain_w_init_48_ver2_stsb

This model is a fine-tuned version of gokuls/bert_12_layer_model_v1_complete_training_new_wt_init_48 on the GLUE STSB dataset. It achieves the following results on the evaluation set:

  • Loss: 2.2509
  • Pearson: 0.1285
  • Spearmanr: 0.1247
  • Combined Score: 0.1266

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
2.3716 1.0 90 2.4198 0.1235 0.0756 0.0995
2.1648 2.0 180 2.4218 0.0592 0.0606 0.0599
2.1915 3.0 270 2.5305 0.1143 0.0959 0.1051
2.1855 4.0 360 2.4912 0.1118 0.0969 0.1043
2.1858 5.0 450 2.3539 0.1130 0.1043 0.1087
2.1818 6.0 540 2.2509 0.1285 0.1247 0.1266
2.2562 7.0 630 2.3302 0.1043 0.0974 0.1009
2.2299 8.0 720 2.3749 0.1984 0.1422 0.1703
2.0676 9.0 810 2.3883 0.1300 0.1329 0.1314
1.926 10.0 900 2.5884 0.1259 0.1233 0.1246
1.7701 11.0 990 2.3776 0.1911 0.2059 0.1985

Framework versions

  • Transformers 4.34.0
  • Pytorch 1.14.0a0+410ce96
  • Datasets 2.14.5
  • Tokenizers 0.14.1
Downloads last month
15
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for gokuls/hBERTv1_new_pretrain_w_init_48_ver2_stsb

Dataset used to train gokuls/hBERTv1_new_pretrain_w_init_48_ver2_stsb

Evaluation results