CoolFace
Modelpublic

pszemraj/xtremedistil-l12-h384-uncased-CoLA

sourceHugging Facemitupdated 9mo agoView on Hugging Face
0likes13downloads
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. -->

xtremedistil-l12-h384-uncased-CoLA

This model is a fine-tuned version of microsoft/xtremedistil-l12-h384-uncased on the GLUE COLA dataset. It achieves the following results on the evaluation set:

  • —Loss: 0.4974
  • —Matthews Correlation: 0.5396

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: 0.0001
  • —trainbatchsize: 128
  • —evalbatchsize: 16
  • —seed: 5559
  • —distributed_type: multi-GPU
  • —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • —lrschedulertype: cosine
  • —lrschedulerwarmup_ratio: 0.03
  • —num_epochs: 16.0
  • —mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation LossMatthews Correlation
0.48221.0670.58930.2621
0.46692.01340.58110.3722
0.30773.02010.61500.4383
0.25944.02680.49740.5396
0.215.03350.55940.5182
0.15266.04020.57150.5150
0.17757.04690.66370.5020
0.16818.05360.69580.5131
0.1249.06030.70570.5154
0.111110.06700.81730.5074
0.133211.07370.82530.5260
0.067312.08040.80860.5180
0.051213.08710.84090.5128
0.045714.09380.87600.4947
0.0415.010050.85220.5103
0.048516.010720.85560.5076

Framework versions

  • —Transformers 4.27.0.dev0
  • —Pytorch 1.13.1+cu117
  • —Datasets 2.8.0
  • —Tokenizers 0.13.1