ast-finetuned-audioset-10-10-0.4593-finetuned-common_voice

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: 0.0088
  • Accuracy: 0.995
  • F1: 0.9950
  • Recall: 0.9950
  • Precision: 0.9951
  • Mcc: 0.9938
  • Auc: 1.0000

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: 8
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 10
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Accuracy F1 Recall Precision Mcc Auc
0.3329 1.0 200 0.1555 0.9525 0.9521 0.9525 0.9566 0.9418 0.9991
0.1007 2.0 400 0.1966 0.9525 0.9512 0.9525 0.9559 0.9420 0.9975
0.0243 3.0 600 0.0619 0.98 0.9799 0.9800 0.9805 0.9752 0.9999
0.0007 4.0 800 0.0194 0.995 0.9950 0.9950 0.9950 0.9938 1.0000
0.0 5.0 1000 0.0182 0.9925 0.9925 0.9925 0.9927 0.9907 1.0000
0.0 6.0 1200 0.0106 0.995 0.9950 0.9950 0.9951 0.9938 1.0000
0.0 7.0 1400 0.0098 0.995 0.9950 0.9950 0.9951 0.9938 1.0000
0.0 8.0 1600 0.0093 0.995 0.9950 0.9950 0.9951 0.9938 1.0000
0.0 9.0 1800 0.0089 0.995 0.9950 0.9950 0.9951 0.9938 1.0000
0.0 10.0 2000 0.0088 0.995 0.9950 0.9950 0.9951 0.9938 1.0000

Framework versions

  • Transformers 4.41.0
  • Pytorch 2.3.0+cu121
  • Datasets 2.19.1
  • Tokenizers 0.19.1
Downloads last month
19
Safetensors
Model size
86.2M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for aydink/ast-finetuned-audioset-10-10-0.4593-finetuned-common_voice

Finetuned
(98)
this model