whisper-small-sid-Oreoluwa

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

• Loss: 0.7471
• Wer: 0.3837
• Cer: 0.0902

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: 64
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 500
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
1.5654	0.4177	500	0.7753	0.4749	0.1239
1.2681	0.8354	1000	0.7143	0.4242	0.1041
1.0191	1.2531	1500	0.6706	0.4050	0.0939
1.0208	1.6708	2000	0.6433	0.4024	0.0969
0.7285	2.0886	2500	0.6675	0.3996	0.0955
0.7945	2.5063	3000	0.6479	0.3900	0.0907
0.7713	2.9240	3500	0.6309	0.3807	0.0889
0.5338	3.3417	4000	0.6756	0.3849	0.0902
0.5898	3.7594	4500	0.6804	0.3817	0.0901
0.3340	4.1771	5000	0.7471	0.3837	0.0902

Framework versions

• Transformers 5.0.0
• Pytorch 2.10.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2