evanmazor/whisper-small-krio

<!-- 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. -->

whisper-small-krio

This model is a fine-tuned version of openai/whisper-small on an unknown dataset. It achieves the following results on the evaluation set:

• —eval_loss: 1.2442
• —eval_wer: 71.1635
• —evalsentencesimilarity_mlm: 0.6126
• —evalsentencesimilarity_e5: 0.9092
• —eval_runtime: 33.6126
• —evalsamplesper_second: 6.099
• —evalstepsper_second: 0.774
• —epoch: 15.8219
• —step: 300

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: 3e-06
• —trainbatchsize: 16
• —evalbatchsize: 8
• —seed: 42
• —gradientaccumulationsteps: 4
• —totaltrainbatch_size: 64
• —optimizer: Use OptimizerNames.ADAMWTORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
• —lrschedulertype: linear
• —lrschedulerwarmup_steps: 100
• —num_epochs: 20
• —mixedprecisiontraining: Native AMP

Framework versions

• —Transformers 4.51.3
• —Pytorch 2.6.0+cu124
• —Datasets 3.5.0
• —Tokenizers 0.21.1