CoolFace
Modelpublic

Gummybear05/wav2vec2-1b-E50_speed

sourceHugging Faceapache-2.0updated 2y agoView on Hugging Face
0likes9downloads
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. -->

wav2vec2-1b-E50_speed

This model is a fine-tuned version of facebook/wav2vec2-xls-r-1b on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.5794
  • Cer: 21.1466

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: 2
  • evalbatchsize: 8
  • seed: 42
  • gradientaccumulationsteps: 8
  • totaltrainbatch_size: 16
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lrschedulertype: linear
  • lrschedulerwarmup_steps: 50
  • num_epochs: 5
  • mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation LossCer
22.6020.25802006.2336100.0
4.85470.51604004.585398.3318
4.61540.77416004.617293.2331
4.47271.03218004.383580.8506
4.28541.290110004.281977.9194
3.9471.548112003.150361.7070
2.0911.806214001.703240.4547
1.2232.064216001.447239.8320
0.88672.322218001.205430.8506
0.74442.580220001.032631.4732
0.65292.838322000.841627.4671
0.53133.096324000.823925.0176
0.41483.354326000.704724.6417
0.36653.612328000.676122.5681
0.34353.870330000.663222.3567
0.28654.128432000.587921.1936
0.23074.386434000.600921.3875
0.21984.644436000.581021.2700
0.1954.902438000.579421.1466

Framework versions

  • Transformers 4.45.2
  • Pytorch 2.3.1.post100
  • Datasets 2.19.1
  • Tokenizers 0.20.1