miosipof/whisper-small-ft-balbus-sep28k-v1.1

miosipof/whisper-tiny-ft-balbus-sep28k-v1.1

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

• Loss: 0.4870
• Accuracy: 0.7719

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-06
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• 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: 0.5
• training_steps: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6991	0.1253	100	0.6929	0.4616
0.686	0.2506	200	0.6816	0.5577
0.6776	0.3759	300	0.6726	0.5631
0.6591	0.5013	400	0.6472	0.6244
0.6317	0.6266	500	0.6115	0.6802
0.5836	0.7519	600	0.5672	0.7104
0.5415	0.8772	700	0.5192	0.7499
0.4856	1.0025	800	0.4999	0.7667
0.4886	1.1278	900	0.4894	0.7715
0.4727	1.2531	1000	0.4870	0.7719

Framework versions

• Transformers 4.48.0
• Pytorch 2.2.0
• Datasets 3.2.0
• Tokenizers 0.21.0