whisper-small-kdn

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: 0.0000
• Wer: 1.1466

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: 0.0004
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 132
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
2.42	2.1277	100	0.2763	38.1356
0.285	4.2553	200	0.1613	21.8843
0.1632	6.3830	300	0.0565	20.5384
0.1055	8.5106	400	0.0679	21.0867
0.126	10.6383	500	0.0546	9.0728
0.0613	12.7660	600	0.0411	11.5155
0.0414	14.8936	700	0.0226	7.5773
0.0282	17.0213	800	0.0190	4.8355
0.0132	19.1489	900	0.0025	2.0439
0.0046	21.2766	1000	0.0007	1.4955
0.0014	23.4043	1100	0.0001	1.7448
0.0001	25.5319	1200	0.0000	1.6451
0.0	27.6596	1300	0.0000	1.1466
0.0	29.7872	1400	0.0000	1.1466

Framework versions

• Transformers 4.45.0.dev0
• Pytorch 2.4.0
• Datasets 2.21.0
• Tokenizers 0.19.1