CoolFace
Modelpublic

gokuls/mobilebert_sa_GLUE_Experiment_logit_kd_stsb_128

sourceHugging Faceapache-2.0updated 4y agoView on Hugging Face
0likes17downloads
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. -->

mobilebertsaGLUEExperimentlogitkdstsb_128

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

  • —Loss: 1.1533
  • —Pearson: 0.0554
  • —Spearmanr: 0.0563
  • —Combined Score: 0.0558

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
  • —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 LossPearsonSpearmanrCombined Score
2.59731.0451.2342-0.0353-0.0325-0.0339
1.09522.0901.17400.04340.04190.0426
1.05813.01351.15330.05540.05630.0558
1.04554.01801.21310.06560.06900.0673
0.97955.02251.38830.08680.08580.0863
0.91976.02701.41410.11810.11480.1165
0.81827.03151.34600.17710.18530.1812
0.67968.03601.15770.22860.23400.2313

Framework versions

  • —Transformers 4.26.0
  • —Pytorch 1.14.0a0+410ce96
  • —Datasets 2.9.0
  • —Tokenizers 0.13.2