CoolFace
Modelpublic

vrclc/Malasar_Medium_DTF

sourceHugging Faceapache-2.0updated 4mo agoView on Hugging Face
0likes10downloads
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. -->

leenag/Malasar_Medium

This model is a fine-tuned version of openai/whisper-medium on the Spoken Bible Corpus: Malasar dataset. It achieves the following results on the evaluation set:

  • Loss: 0.5907
  • Wer: 50.2570

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: 1e-05
  • trainbatchsize: 32
  • evalbatchsize: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lrschedulertype: linear
  • lrschedulerwarmup_steps: 500
  • training_steps: 2000
  • mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation LossWer
0.064611.36362500.336955.6254
0.010422.72735000.444552.3130
0.001234.09097500.489050.1428
0.000245.454510000.524050.3712
0.000256.818212500.548850.1999
0.000168.181815000.569550.3712
0.000179.545517500.584450.1999
0.000190.909120000.590750.2570

Framework versions

  • Transformers 4.41.2
  • Pytorch 2.0.1+cu117
  • Datasets 2.16.0
  • Tokenizers 0.19.1

Citation

bibtex
@misc{multistage2024,
      title={Multistage Fine-tuning Strategies for Automatic Speech Recognition in Low-resource Languages}, 
      author={Leena G Pillai, Kavya Manohar, Basil K Raju, Elizabeth Sherly},
      year={2024},
      eprint={2411.04573},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2411.04573}, 
}