CoolFace
Modelpublic

infinitejoy/wav2vec2-large-xls-r-300m-bashkir

sourceHugging Faceapache-2.0updated 5y agoView on Hugging Face
0likes108downloads
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-xls-r-300m-bashkir

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the MOZILLA-FOUNDATION/COMMONVOICE7_0 - BA dataset. It achieves the following results on the evaluation set:

  • —Loss: 0.1892
  • —Wer: 0.2421

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: 32
  • —evalbatchsize: 32
  • —seed: 42
  • —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • —lrschedulertype: linear
  • —lrschedulerwarmup_steps: 2000
  • —num_epochs: 10.0
  • —mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation LossWer
1.47920.520000.45980.5404
1.4491.040000.46500.5610
1.37421.4960000.40010.4977
1.33751.9980000.39160.4894
1.29612.49100000.36410.4569
1.27142.99120000.34910.4488
1.23993.48140000.31510.3986
1.20673.98160000.30810.3923
1.18424.48180000.28750.3703
1.16444.98200000.28400.3670
1.1615.48220000.27900.3597
1.13035.97240000.25520.3272
1.08746.47260000.24050.3142
1.06136.97280000.23520.3055
1.04987.47300000.22490.2910
1.0217.96320000.21180.2752
1.00028.46340000.20460.2662
0.97628.96360000.19690.2530
0.95689.46380000.19170.2449
0.9539.96400000.18930.2425

Framework versions

  • —Transformers 4.16.0.dev0
  • —Pytorch 1.10.1+cu102
  • —Datasets 1.17.1.dev0
  • —Tokenizers 0.11.0