Modelpublic

IoanaLivia/real_data_1_h_synth_standard_A_horoscope-whisper-small

sourceHugging Faceapache-2.0updated 1y agoView on Hugging Face

0likes2downloads

Model Card

IoanaLivia/realdata1hsynthstandardA_horoscope-whisper-small

This model is a fine-tuned version of openai/whisper-small on the IoanaLivia/realdata1hsynthstandardA_horoscope dataset. It achieves the following results on the evaluation set:

—Loss: 0.3404
—Wer: 18.4631

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: 8
—evalbatchsize: 4
—seed: 42
—gradientaccumulationsteps: 2
—totaltrainbatch_size: 16
—optimizer: Use OptimizerNames.ADAMWTORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: linear
—num_epochs: 15
—mixedprecisiontraining: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	0	0	0.4170	30.1014
0.3105	1.0	49	0.2822	18.6350
0.1247	2.0	98	0.2664	18.8241
0.0589	3.0	147	0.2849	19.1164
0.0283	4.0	196	0.3118	18.6350
0.0143	5.0	245	0.3264	18.5319
0.0077	6.0	294	0.3404	18.4631
0.0058	7.0	343	0.3490	18.9273
0.0041	8.0	392	0.3579	18.8585
0.0032	9.0	441	0.3654	19.0132
0.0028	10.0	490	0.3694	19.0476
0.0025	11.0	539	0.3726	19.1164
0.0023	12.0	588	0.3753	19.1851
0.0022	13.0	637	0.3770	19.1680
0.0021	14.0	686	0.3779	19.1680
0.0021	14.7010	720	0.3780	19.2023

Framework versions

—Transformers 4.51.3
—Pytorch 2.6.0+cu124
—Datasets 3.6.0
—Tokenizers 0.21.1