ixxan/wav2vec2-large-mms-1b-uyghur-latin

<!-- 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-mms-1b-uyghur-latin

This model is a fine-tuned version of facebook/mms-1b-all on the None dataset. It achieves the following best results on the evaluation set:

• —Best Wer: 30.8949%
• —Best Cer: 5.9823 %

Training procedure

Finetuning code avaiblable in https://github.com/ixxan/ug-speech

Training hyperparameters

The following hyperparameters were used during training:

• —learning_rate: 0.001
• —trainbatchsize: 8
• —evalbatchsize: 8
• —seed: 42
• —gradientaccumulationsteps: 4
• —totaltrainbatch_size: 32
• —optimizer: Use adamwtorch with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
• —lrschedulertype: linear
• —lrschedulerwarmup_steps: 500
• —num_epochs: 4
• —mixedprecisiontraining: Native AMP

Training results

Framework versions

• —Transformers 4.46.2
• —Pytorch 2.5.1+cu121
• —Datasets 3.1.0
• —Tokenizers 0.20.3