ast-finetuned-audioset-10-10-0.4593-finetuned-rfcx-reprod

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: 0.4964
• Accuracy: 0.9475

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: 5e-05
• train_batch_size: 12
• eval_batch_size: 12
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.95,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: reduce_lr_on_plateau
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0	0	0.8397	0.5805
0.1754	1.8896	5000	0.2016	0.9368
0.1259	3.7793	10000	0.2081	0.9481
0.0662	5.6689	15000	0.2144	0.9515
0.0507	7.5586	20000	0.3208	0.9461
0.0367	9.4482	25000	0.3234	0.9518
0.027	11.3379	30000	0.3243	0.9476
0.0123	13.2275	35000	0.3838	0.9470
0.0001	15.1172	40000	0.4480	0.9521

Framework versions

• Transformers 4.48.0
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0