CoolFace
Modelpublic

dracoglacius/wav2vec2-large-960h-lv60-self-with-wikipedia-lm-timit

sourceHugging Faceupdated 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-large-960h-lv60-self-with-wikipedia-lm-timit

This model is a fine-tuned version of gxbag/wav2vec2-large-960h-lv60-self-with-wikipedia-lm on the None dataset. It achieves the following results on the evaluation set:

  • —Loss: 0.0889
  • —Wer: 0.4976

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: 0.0003
  • —trainbatchsize: 16
  • —evalbatchsize: 8
  • —seed: 42
  • —gradientaccumulationsteps: 2
  • —totaltrainbatch_size: 32
  • —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • —lrschedulertype: linear
  • —lrschedulerwarmup_steps: 1000
  • —num_epochs: 30
  • —mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation LossWer
7.79112.022503.08961.0
1.38544.035000.07040.5052
0.19266.057500.06780.5010
0.14728.0610000.07940.5157
0.132610.0812500.09370.5031
0.10412.115000.08590.5055
0.075414.1117500.09030.5031
0.062416.1320000.09270.5034
0.059418.1422500.09290.5016
0.05720.1625000.08730.5039
0.047622.1827500.09740.5055
0.038224.1930000.08860.5003
0.032926.2132500.08320.4987
0.03228.2235000.08890.4976

Framework versions

  • —Transformers 4.23.0.dev0
  • —Pytorch 1.13.0.dev20220624+cu113
  • —Datasets 2.5.2.dev0
  • —Tokenizers 0.12.1