wa976's picture
update model card README.md
5ab9d8a
metadata
license: bsd-3-clause
tags:
  - generated_from_trainer
metrics:
  - accuracy
model-index:
  - name: ast_15-finetuned-ICBHI
    results: []

ast_15-finetuned-ICBHI

This model is a fine-tuned version of MIT/ast-finetuned-audioset-10-10-0.4593 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 1.1688
  • Accuracy: 0.5397
  • Sensitivity: 0.2727
  • Specificity: 0.7389
  • Score: 0.5058

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: 3e-05
  • train_batch_size: 4
  • eval_batch_size: 4
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 16
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 15

Training results

Training Loss Epoch Step Validation Loss Accuracy Sensitivity Specificity Score
0.7488 1.0 259 1.1831 0.5241 0.3551 0.6502 0.5027
0.7831 2.0 518 1.1688 0.5397 0.2727 0.7389 0.5058
0.7471 3.0 777 1.1593 0.5198 0.3772 0.6261 0.5017
0.5336 4.0 1036 1.4082 0.5281 0.3152 0.6869 0.5011
0.3833 5.0 1295 2.0232 0.4838 0.3840 0.5583 0.4712
0.1721 6.0 1554 2.5558 0.4893 0.3534 0.5906 0.4720
0.2745 7.0 1813 3.3175 0.4900 0.3917 0.5634 0.4775
0.0596 8.0 2072 3.6548 0.5143 0.3628 0.6274 0.4951
0.0034 9.0 2331 3.9119 0.5082 0.4053 0.5849 0.4951
0.0008 10.0 2590 4.3407 0.4875 0.4562 0.5108 0.4835
0.0 11.0 2849 4.1927 0.5136 0.3636 0.6255 0.4946
0.0 12.0 3108 4.2227 0.5111 0.3645 0.6204 0.4924
0.0 13.0 3367 4.2399 0.5114 0.3653 0.6204 0.4929
0.0 14.0 3626 4.2521 0.5114 0.3662 0.6198 0.4930
0.0 15.0 3885 4.2556 0.5114 0.3662 0.6198 0.4930

Framework versions

  • Transformers 4.29.2
  • Pytorch 2.0.0+cu118
  • Datasets 2.12.0
  • Tokenizers 0.13.3