wa976's picture
update model card README.md
d191743
metadata
license: bsd-3-clause
tags:
  - generated_from_trainer
metrics:
  - accuracy
model-index:
  - name: ast_binary_7-finetuned-ICBHI
    results: []

ast_binary_7-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: 0.7561
  • Accuracy: 0.5764
  • Sensitivity: 0.6185
  • Specificity: 0.5450
  • Score: 0.5818

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: 7

Training results

Training Loss Epoch Step Validation Loss Accuracy Sensitivity Specificity Score
0.6321 1.0 259 0.7561 0.5764 0.6185 0.5450 0.5818
0.5672 2.0 518 0.8579 0.5626 0.6015 0.5336 0.5676
0.5443 3.0 777 1.0517 0.5074 0.8275 0.2687 0.5481
0.5075 4.0 1036 0.9977 0.5358 0.7638 0.3657 0.5647
0.4912 5.0 1295 1.2474 0.4969 0.8539 0.2307 0.5423
0.4331 6.0 1554 1.0732 0.5376 0.7077 0.4106 0.5592
0.4368 7.0 1813 1.0947 0.5405 0.7230 0.4043 0.5637

Framework versions

  • Transformers 4.29.2
  • Pytorch 2.0.1+cu118
  • Datasets 2.12.0
  • Tokenizers 0.13.3