bird-call-classification

This model is a fine-tuned version of MIT/ast-finetuned-audioset-10-10-0.4593 on the kayalvizhi42/bird_calls. It achieves the following results on the evaluation set:

• Loss: 0.0735
• Accuracy: 0.9799
• Precision: 0.9579
• Recall: 1.0
• F1: 0.9785

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: 0.0002
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• 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
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
No log	1.0	50	0.2127	0.9497	0.9175	0.9780	0.9468
No log	2.0	100	0.1252	0.9698	0.9474	0.9890	0.9677
No log	3.0	150	0.1037	0.9749	0.9479	1.0	0.9733
No log	4.0	200	0.0947	0.9698	0.9381	1.0	0.9681
No log	5.0	250	0.0850	0.9799	0.9579	1.0	0.9785
No log	6.0	300	0.0802	0.9799	0.9579	1.0	0.9785
No log	7.0	350	0.0789	0.9799	0.9579	1.0	0.9785
No log	8.0	400	0.0769	0.9799	0.9579	1.0	0.9785
No log	9.0	450	0.0736	0.9799	0.9579	1.0	0.9785
0.1077	10.0	500	0.0735	0.9799	0.9579	1.0	0.9785

Framework versions

• Transformers 4.48.1
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0