whisper-small-ja – AI Model by toshiouchiyama | AlphaNeural AI

You can deploy this model and start earning money today!

toshiouchiyama/
whisper-small-ja

like

0

automatic-speech-recognition

generated_from_trainer

endpoints_compatible

Views

No views yet

whisper-small-ja

This model is a fine-tuned version of openai/whisper-small on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.3967
Wer: 18.3755

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
train_batch_size: 8
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 5
training_steps: 200

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	0.3	10	1.1627	26.0985
No log	0.61	20	0.7416	900.3995
1.2431	0.91	30	0.6344	60.3196
1.2431	1.21	40	0.5944	20.2397
0.5462	1.52	50	0.5341	19.3076
0.5462	1.82	60	0.4953	18.5087
0.5462	2.12	70	0.4715	19.9734
0.3259	2.42	80	0.4469	18.2423
0.3259	2.73	90	0.4246	19.7071
0.1986	3.03	100	0.4076	19.0413
0.1986	3.33	110	0.3949	17.7097
0.1986	3.64	120	0.4008	20.5060
0.1101	3.94	130	0.3892	18.3755
0.1101	4.24	140	0.3873	18.3755
0.0695	4.55	150	0.3930	19.7071
0.0695	4.85	160	0.3857	18.1092
0.0695	5.15	170	0.3861	19.0413
0.0467	5.45	180	0.3913	18.5087
0.0467	5.76	190	0.3963	18.7750
0.0346	6.06	200	0.3967	18.3755

Framework versions

Transformers 4.26.0.dev0
Pytorch 1.13.1+cpu
Datasets 2.8.0
Tokenizers 0.13.2