Rest970828's picture
Model save
863c158 verified
metadata
license: apache-2.0
base_model: facebook/wav2vec2-base-960h
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - f1
model-index:
  - name: wav2vec2-base-960h-finetuned-ks
    results: []

wav2vec2-base-960h-finetuned-ks

This model is a fine-tuned version of facebook/wav2vec2-base-960h on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.9308
  • Accuracy: 0.7752
  • F1: 0.7749

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: 32
  • eval_batch_size: 32
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 128
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 50

Training results

Training Loss Epoch Step Validation Loss Accuracy F1
1.3739 0.99 35 1.3646 0.3654 0.2858
1.3444 2.0 71 1.3366 0.3833 0.3172
1.3193 2.99 106 1.2654 0.4324 0.3350
1.2447 4.0 142 1.2093 0.4649 0.3611
1.2087 4.99 177 1.2030 0.4582 0.3714
1.1539 6.0 213 1.1419 0.4920 0.4317
1.0795 6.99 248 1.1794 0.4721 0.4207
1.0525 8.0 284 1.0922 0.5020 0.4684
1.0615 8.99 319 1.0459 0.5471 0.5158
0.9381 10.0 355 1.0080 0.5656 0.5464
0.8945 10.99 390 1.1166 0.5378 0.5108
0.8497 12.0 426 1.0068 0.5855 0.5772
0.7729 12.99 461 1.1214 0.5517 0.5406
0.6984 14.0 497 1.0416 0.5889 0.5729
0.6856 14.99 532 1.0135 0.6180 0.6185
0.6095 16.0 568 1.0088 0.6320 0.6299
0.5899 16.99 603 0.9208 0.6585 0.6612
0.5922 18.0 639 0.8657 0.6757 0.6749
0.537 18.99 674 0.8910 0.6850 0.6892
0.4767 20.0 710 1.0544 0.6525 0.6499
0.4864 20.99 745 0.8024 0.7255 0.7232
0.3546 22.0 781 0.8628 0.7168 0.7205
0.3567 22.99 816 0.8921 0.7168 0.7177
0.381 24.0 852 0.9130 0.7069 0.7081
0.3031 24.99 887 1.0026 0.7023 0.7039
0.412 26.0 923 0.8413 0.7420 0.7430
0.3175 26.99 958 0.8705 0.7294 0.7335
0.2581 28.0 994 0.8628 0.7387 0.7431
0.328 28.99 1029 0.9022 0.7414 0.7417
0.263 30.0 1065 0.9787 0.7248 0.7251
0.249 30.99 1100 0.8658 0.7454 0.7481
0.2242 32.0 1136 0.9386 0.7354 0.7380
0.2848 32.99 1171 0.8553 0.7633 0.7639
0.2457 34.0 1207 0.8789 0.7692 0.7674
0.1557 34.99 1242 0.8542 0.7553 0.7594
0.169 36.0 1278 0.9132 0.7573 0.7600
0.171 36.99 1313 0.9550 0.7467 0.7481
0.2209 38.0 1349 0.9843 0.7407 0.7408
0.1674 38.99 1384 0.9523 0.7460 0.7468
0.1998 40.0 1420 0.8683 0.7686 0.7697
0.1101 40.99 1455 1.0123 0.7354 0.7370
0.1466 42.0 1491 0.9332 0.7633 0.7651
0.1376 42.99 1526 0.9193 0.7739 0.7743
0.0939 44.0 1562 0.9234 0.7626 0.7634
0.1333 44.99 1597 0.9308 0.7752 0.7749
0.1183 46.0 1633 0.9375 0.7706 0.7712
0.1031 46.99 1668 0.9298 0.7739 0.7750
0.1154 48.0 1704 0.9373 0.7739 0.7745
0.1317 48.99 1739 0.9611 0.7646 0.7654
0.1132 49.3 1750 0.9606 0.7626 0.7635

Framework versions

  • Transformers 4.36.2
  • Pytorch 2.1.1+cu121
  • Datasets 2.16.1
  • Tokenizers 0.15.0