beewatch_ast_3class

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: 0.0029
• Accuracy: 1.0
• Macro Precision: 1.0
• Macro Recall: 1.0
• Macro F1: 1.0

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: 16
• eval_batch_size: 16
• seed: 42
• 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: 10
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Macro Precision	Macro Recall	Macro F1
0.2254	1.0	48	0.1543	0.9509	0.9386	0.9490	0.9429
0.0857	2.0	96	0.0421	0.9877	0.9810	0.9897	0.9850
0.0213	3.0	144	0.0253	0.9877	0.9900	0.9798	0.9845
0.0009	4.0	192	0.0066	1.0	1.0	1.0	1.0
0.0004	5.0	240	0.0062	1.0	1.0	1.0	1.0
0.0002	6.0	288	0.0036	1.0	1.0	1.0	1.0
0.0002	7.0	336	0.0035	1.0	1.0	1.0	1.0
0.0002	8.0	384	0.0035	1.0	1.0	1.0	1.0
0.0002	9.0	432	0.0032	1.0	1.0	1.0	1.0
0.0001	10.0	480	0.0031	1.0	1.0	1.0	1.0
0.0001	11.0	528	0.0031	1.0	1.0	1.0	1.0
0.0001	12.0	576	0.0031	1.0	1.0	1.0	1.0
0.0001	13.0	624	0.0030	1.0	1.0	1.0	1.0
0.0001	14.0	672	0.0029	1.0	1.0	1.0	1.0
0.0001	15.0	720	0.0029	1.0	1.0	1.0	1.0

Framework versions

• Transformers 5.16.1
• Pytorch 2.11.0+cu128
• Datasets 4.8.5
• Tokenizers 0.23.1