CoolFace
Modelpublic

OverloadedOperator/tokomni-whisper-i3

sourceHugging Faceapache-2.0updated 2y 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. -->

TKWhisperASR

This model is a fine-tuned version of openai/whisper-tiny on the KAN Hebrew Speech dataset. It achieves the following results on the evaluation set:

  • —Loss: 5.4256

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: 2
  • —evalbatchsize: 2
  • —seed: 42
  • —gradientaccumulationsteps: 4
  • —totaltrainbatch_size: 8
  • —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • —lrschedulertype: linear
  • —lrschedulerwarmup_steps: 50
  • —num_epochs: 30
  • —mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation Loss
No log1.016.4006
No log2.026.4006
No log3.036.4006
No log4.046.4006
No log5.056.4006
No log6.066.3966
No log7.076.3910
No log8.086.3844
No log9.096.3857
No log10.0106.3694
No log11.0116.3381
No log12.0126.3115
No log13.0136.2788
No log14.0146.2411
No log15.0156.1932
No log16.0166.1552
No log17.0176.1146
No log18.0186.0711
No log19.0196.0122
No log20.0205.9643
No log21.0215.9193
No log22.0225.8783
No log23.0235.8258
No log24.0245.7742
6.730625.0255.7174
6.730626.0265.6629
6.730627.0275.6041
6.730628.0285.5453
6.730629.0295.4896
6.730630.0305.4256

Framework versions

  • —PEFT 0.9.0
  • —Transformers 4.38.2
  • —Pytorch 2.2.1+cu121
  • —Datasets 2.18.0
  • —Tokenizers 0.15.0