CoolFace
Modelpublic

gokuls/hBERTv1_new_pretrain_48_sst2

sourceHugging Faceupdated 3y agoView on Hugging Face
0likes22downloads
Model Card

<!-- 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. -->

hBERTv1newpretrain48sst2

This model is a fine-tuned version of gokuls/bert_12_layer_model_v1_complete_training_new_48 on the GLUE SST2 dataset. It achieves the following results on the evaluation set:

  • —Loss: 0.4843
  • —Accuracy: 0.8085

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
  • —trainbatchsize: 128
  • —evalbatchsize: 128
  • —seed: 10
  • —distributed_type: multi-GPU
  • —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • —lrschedulertype: linear
  • —num_epochs: 50

Training results

Training LossEpochStepValidation LossAccuracy
0.38941.05270.48430.8085
0.23532.010540.60850.7947
0.19173.015810.55470.7901
0.16534.021080.64570.8050
0.14535.026350.56100.7936
0.12966.031620.64570.7867

Framework versions

  • —Transformers 4.29.2
  • —Pytorch 1.14.0a0+410ce96
  • —Datasets 2.12.0
  • —Tokenizers 0.13.3