Modelpublic

Eimhin03/output_model_whisper_base_shunya_ideal_data_augmentation_enabled_attepmt2

sourceHugging Faceapache-2.0updated 7mo agoView on Hugging Face

0likes6downloads

Model Card

outputmodelwhisperbaseshunyaidealdataaugmentationenabled_attepmt2

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

—Loss: nan
—Wer: 100.0

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: 0.001
—trainbatchsize: 2
—evalbatchsize: 4
—seed: 42
—optimizer: Use OptimizerNames.ADAMWTORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: linear
—lrschedulerwarmup_steps: 500
—training_steps: 100000
—mixedprecisiontraining: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
2.2403	3.9062	5000	2.6294	114.8722
0.7554	7.8125	10000	1.8172	103.7218
0.0	11.7188	15000	nan	100.0
0.0	15.625	20000	nan	100.0
0.0	19.5312	25000	nan	100.0
0.0	23.4375	30000	nan	100.0
0.0	27.3438	35000	nan	100.0
0.0	31.25	40000	nan	100.0
0.0	35.1562	45000	nan	100.0
0.0	39.0625	50000	nan	100.0
0.0	42.9688	55000	nan	100.0
0.0	46.875	60000	nan	100.0
0.0	50.7812	65000	nan	100.0
0.0	54.6875	70000	nan	100.0
0.0	58.5938	75000	nan	100.0
0.0	62.5	80000	nan	100.0
0.0	66.4062	85000	nan	100.0
0.0	70.3125	90000	nan	100.0
0.0	74.2188	95000	nan	100.0
0.0	78.125	100000	nan	100.0

Framework versions

—Transformers 5.3.0.dev0
—Pytorch 2.9.0+cu126
—Datasets 4.0.0
—Tokenizers 0.22.2