whisper-large-v3-finetuned-5

This model is a fine-tuned version of openai/whisper-large-v3 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3599
• Wer: 50.9574

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-08
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.934	1.0	7532	0.9560	16.8450
0.8091	2.0	15064	0.5512	14.8480
1.8408	3.0	22596	0.4653	14.6010
0.1874	4.0	30128	0.4304	14.4735
0.0178	5.0	37660	0.3972	14.2372
1.2841	6.0	45192	0.3756	53.3820
0.0041	7.0	52724	0.3663	49.1462
0.3521	8.0	60256	0.3628	50.0624
0.0849	9.0	67788	0.3604	51.1061
1.2129	10.0	75320	0.3599	50.9574

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2