CoolFace
Modelpublic

MelanieKoe/w2v2-base-pretrained_lr1e-4_at0.8_da0.4

sourceHugging Faceapache-2.0updated 2y agoView on Hugging Face
0likes8downloads
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. -->

w2v2-base-pretrainedlr1e-4at0.8_da0.4

This model is a fine-tuned version of facebook/wav2vec2-base on the None dataset. It achieves the following results on the evaluation set:

  • —Loss: 2.1115
  • —Wer: 0.1952

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: 32
  • —evalbatchsize: 8
  • —seed: 42
  • —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • —lrschedulertype: linear
  • —lrschedulerwarmup_steps: 1000
  • —training_steps: 3500
  • —mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation LossWer
21.215713.162504.16341.0
3.233726.325003.12311.0
3.057539.477503.04661.0
2.073952.6310001.06770.6284
0.175865.7912501.37110.3170
0.067578.9515001.65210.2268
0.035592.1117501.73130.2332
0.0209105.2620001.97200.2114
0.0162118.4222501.75690.2085
0.0099131.5825002.16230.1944
0.0071144.7427502.20670.1922
0.0066157.8930002.12460.1944
0.0059171.0532502.14840.1922
0.0045184.2135002.11150.1952

Framework versions

  • —Transformers 4.35.0
  • —Pytorch 2.0.0
  • —Datasets 2.14.6
  • —Tokenizers 0.14.1