whisper-a-nomimose-ag

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: 0.0208
• Wer: 9.7345

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: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 132
• num_epochs: 12
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.0018	0.9217	100	0.1203	22.7139
0.2226	1.8387	200	0.0800	20.8702
0.0921	2.7558	300	0.0724	24.7050
0.0509	3.6728	400	0.0728	22.2714
0.0436	4.5899	500	0.0298	18.1416
0.0195	5.5069	600	0.0498	22.2714
0.0193	6.4240	700	0.0369	18.7316
0.0156	7.3410	800	0.0216	16.5192
0.0117	8.2581	900	0.0242	14.0118
0.0069	9.1751	1000	0.0191	10.2507
0.0043	10.0922	1100	0.0195	8.7758
0.0028	11.0092	1200	0.0208	9.7345

Framework versions

• Transformers 4.47.0.dev0
• Pytorch 2.4.0
• Datasets 3.0.1
• Tokenizers 0.20.0