whisper-tiny-aug-19-mar-v1

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

• Loss: 0.2310
• Wer: 103.8677

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• lr_scheduler_warmup_steps: 1000
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.3938	1.0	381	1.1161	100.3684
0.7824	2.0	762	0.5599	103.7396
0.4849	3.0	1143	0.4141	101.4814
0.3804	4.0	1524	0.3526	102.8507
0.322	5.0	1905	0.3112	100.9449
0.2803	6.0	2286	0.2866	107.6073
0.2501	7.0	2667	0.2667	104.9808
0.2258	8.0	3048	0.2520	101.4414
0.2056	9.0	3429	0.2406	104.1480
0.1881	10.0	3810	0.2310	103.8677

Framework versions

• Transformers 4.49.0
• Pytorch 2.4.1+cu121
• Datasets 3.4.1
• Tokenizers 0.21.1