red_round_button_small_click_detector_n100

This model is a fine-tuned version of MIT/ast-finetuned-audioset-10-10-0.4593 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0899
• Recall Pressed: 0.8984
• Precision Pressed: 0.9659
• F1 Pressed: 0.9309
• False Negative Rate: 0.1016

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: 82
• eval_batch_size: 82
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Recall Pressed	Precision Pressed	F1 Pressed	False Negative Rate
0.0195	1.0	117	0.0575	0.9746	0.8797	0.9247	0.0254
0.0052	2.0	234	0.0483	0.9079	0.9286	0.9181	0.0921
0.0093	3.0	351	0.0832	0.8794	0.9928	0.9327	0.1206
0.0130	4.0	468	0.0607	0.9016	0.9498	0.9251	0.0984
0.0086	5.0	585	0.0601	0.8921	0.9623	0.9259	0.1079
0.0046	6.0	702	0.0888	0.8889	0.9859	0.9349	0.1111
0.0026	7.0	819	0.0763	0.9048	0.9727	0.9375	0.0952
0.0023	8.0	936	0.0853	0.8984	0.9692	0.9325	0.1016
0.0002	9.0	1053	0.0920	0.8921	0.9723	0.9305	0.1079
0.0024	10.0	1170	0.0901	0.8984	0.9659	0.9309	0.1016

Framework versions

• Transformers 5.11.0
• Pytorch 2.7.1+cu118
• Datasets 5.0.0
• Tokenizers 0.22.2