abbenedekwhisper-base.en-finetuning2-D3K

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

• Loss: 4.7781
• Cer: 64.7190
• Wer: 119.5364
• Ser: 100.0
• Cer Clean: 3.5058
• Wer Clean: 6.2914
• Ser Clean: 7.0175

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: 5e-08
• train_batch_size: 16
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 10
• training_steps: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer	Wer	Ser	Cer Clean	Wer Clean	Ser Clean
7.5369	0.53	100	6.7220	63.7730	128.1457	100.0	4.1180	6.9536	8.7719
7.0363	1.06	200	6.1829	65.0529	123.8411	100.0	3.2832	5.6291	7.0175
6.417	1.6	300	5.7959	64.1625	121.1921	100.0	3.2832	5.6291	7.0175
6.0146	2.13	400	5.4587	64.7746	121.8543	100.0	3.6728	6.6225	7.8947
5.6687	2.66	500	5.2287	65.3311	120.5298	100.0	3.7284	6.6225	7.8947
5.3902	3.19	600	5.0691	65.1085	121.1921	100.0	3.5615	6.2914	7.0175
5.2512	3.72	700	4.9358	64.7190	120.1987	100.0	3.2832	5.9603	6.1404
5.1258	4.26	800	4.8451	64.7190	119.5364	100.0	3.5058	6.2914	7.0175
5.0472	4.79	900	4.7950	64.7190	119.5364	100.0	3.5058	6.2914	7.0175
4.9871	5.32	1000	4.7781	64.7190	119.5364	100.0	3.5058	6.2914	7.0175

Framework versions

• Transformers 4.39.3
• Pytorch 2.2.2+cu121
• Datasets 2.14.5
• Tokenizers 0.15.2