whisper-tiny-finetuned-bmd-mx30-shfl-20241112_105222

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

• Loss: 1.8366
• Accuracy: 0.4706
• F1: 0.4755

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: 0.0001
• train_batch_size: 16
• eval_batch_size: 16
• seed: 1968
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• 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_ratio: 0.1
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	0.8571	3	1.1053	0.2647	0.1108
No log	2.0	7	1.1082	0.2941	0.1683
1.0599	2.8571	10	1.0908	0.3529	0.2644
1.0599	4.0	14	1.0633	0.3824	0.3022
1.0599	4.8571	17	0.9311	0.4706	0.4514
0.6821	6.0	21	1.1443	0.4412	0.4371
0.6821	6.8571	24	1.1714	0.5	0.5002
0.6821	8.0	28	1.2322	0.5294	0.5345
0.193	8.8571	31	1.5522	0.4412	0.4147
0.193	10.0	35	1.7296	0.4706	0.4540
0.193	10.8571	38	1.7856	0.4412	0.4425
0.035	12.0	42	1.8251	0.4412	0.4423
0.035	12.8571	45	1.8366	0.4706	0.4755

Framework versions

• Transformers 4.46.2
• Pytorch 2.5.0+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3