Edit model card

mobilebert_sa_GLUE_Experiment_logit_kd_pretrain_sst2

This model is a fine-tuned version of gokuls/mobilebert_sa_pre-training-complete on the GLUE SST2 dataset. It achieves the following results on the evaluation set:

  • Loss: 0.2364
  • Accuracy: 0.9266

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: 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 Accuracy
0.4176 1.0 527 0.2978 0.9197
0.1807 2.0 1054 0.2951 0.9174
0.1163 3.0 1581 0.2749 0.9186
0.0862 4.0 2108 0.2988 0.9083
0.0695 5.0 2635 0.2760 0.9174
0.0598 6.0 3162 0.2695 0.9151
0.0525 7.0 3689 0.2723 0.9255
0.0464 8.0 4216 0.2430 0.9243
0.0422 9.0 4743 0.2814 0.9243
0.0395 10.0 5270 0.2464 0.9163
0.0357 11.0 5797 0.2390 0.9197
0.0341 12.0 6324 0.2713 0.9197
0.0328 13.0 6851 0.2685 0.9220
0.0315 14.0 7378 0.2585 0.9186
0.0296 15.0 7905 0.2367 0.9220
0.0283 16.0 8432 0.2560 0.9186
0.0277 17.0 8959 0.2635 0.9174
0.0269 18.0 9486 0.2364 0.9266
0.026 19.0 10013 0.2749 0.9209
0.0252 20.0 10540 0.2507 0.9174
0.0248 21.0 11067 0.2769 0.9163
0.0248 22.0 11594 0.2543 0.9220
0.024 23.0 12121 0.2677 0.9209

Framework versions

  • Transformers 4.26.0
  • Pytorch 1.14.0a0+410ce96
  • Datasets 2.9.0
  • Tokenizers 0.13.2
Downloads last month
7

Dataset used to train gokuls/mobilebert_sa_GLUE_Experiment_logit_kd_pretrain_sst2

Evaluation results