CoolFace
Modelpublic

Jungwoo4021/wav2vec2-base-ks-padpt400

sourceHugging Faceapache-2.0updated 4y agoView on Hugging Face
0likes18downloads
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. -->

wav2vec2-base-ks-padpt400

This model is a fine-tuned version of facebook/wav2vec2-base on the superb dataset. It achieves the following results on the evaluation set:

  • —Loss: 1.2218
  • —Accuracy: 0.6343

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: 0.003
  • —trainbatchsize: 256
  • —evalbatchsize: 256
  • —seed: 0
  • —gradientaccumulationsteps: 4
  • —totaltrainbatch_size: 1024
  • —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • —lrschedulertype: linear
  • —lrschedulerwarmup_ratio: 0.1
  • —num_epochs: 10.0
  • —mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation LossAccuracy
1.29481.0501.65270.6108
0.88612.01001.26530.6130
0.78093.01501.26150.5924
0.73644.02001.22180.6343
0.69445.02501.21370.6324
0.68176.03001.28220.5930
0.66017.03501.32920.5599
0.64648.04001.27440.5869
0.6539.04501.39160.5272
0.63310.05001.33440.5606

Framework versions

  • —Transformers 4.22.0.dev0
  • —Pytorch 1.11.0+cu115
  • —Datasets 2.4.0
  • —Tokenizers 0.12.1