Edit model card

vakyansh-wav2vec2-kannada-knm-560-audio-abuse-feature

This model is a fine-tuned version of Harveenchadha/vakyansh-wav2vec2-kannada-knm-560 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.6403
  • Accuracy: 0.7100
  • Macro F1-score: 0.6596

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: 2e-05
  • train_batch_size: 16
  • eval_batch_size: 16
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 64
  • 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 Macro F1-score
6.6756 0.77 10 6.6487 0.0 0.0
6.6336 1.54 20 6.5448 0.5474 0.0647
6.4999 2.31 30 6.3245 0.6585 0.3971
6.2688 3.08 40 6.0120 0.6585 0.3971
6.0598 3.85 50 5.7401 0.6585 0.3971
5.7739 4.62 60 5.4859 0.6585 0.3971
5.5736 5.38 70 5.2443 0.6585 0.3971
5.3092 6.15 80 5.0361 0.6585 0.3971
5.1088 6.92 90 4.8282 0.6585 0.3971
4.9566 7.69 100 4.6295 0.6585 0.3971
4.7528 8.46 110 4.4350 0.6585 0.3971
4.6942 9.23 120 4.2479 0.6585 0.3971
4.4164 10.0 130 4.0578 0.6585 0.3971
4.1989 10.77 140 3.8571 0.6585 0.3971
4.0312 11.54 150 3.6581 0.6585 0.3971
3.8758 12.31 160 3.4561 0.6585 0.3971
3.7026 13.08 170 3.2569 0.6585 0.3971
3.4173 13.85 180 3.0592 0.6585 0.3971
3.2018 14.62 190 2.8633 0.6585 0.3971
3.1789 15.38 200 2.6746 0.6585 0.3971
2.8636 16.15 210 2.4860 0.6585 0.3971
2.6381 16.92 220 2.3059 0.6585 0.3971
2.5071 17.69 230 2.1303 0.6585 0.3971
2.2478 18.46 240 1.9669 0.6585 0.3971
2.2718 19.23 250 1.8162 0.6585 0.3971
2.0259 20.0 260 1.6750 0.6585 0.3971
1.8823 20.77 270 1.5460 0.6585 0.3971
1.6591 21.54 280 1.4290 0.6585 0.3971
1.5646 22.31 290 1.3213 0.6585 0.3971
1.487 23.08 300 1.2263 0.6585 0.3971
1.3681 23.85 310 1.1424 0.6585 0.3971
1.2941 24.62 320 1.0696 0.6585 0.3971
1.1374 25.38 330 1.0059 0.6585 0.3971
1.0881 26.15 340 0.9470 0.6585 0.3971
0.9892 26.92 350 0.8987 0.6585 0.3971
1.0156 27.69 360 0.8547 0.6585 0.3971
0.9592 28.46 370 0.8181 0.6585 0.3971
0.937 29.23 380 0.7861 0.6585 0.3971
0.8938 30.0 390 0.7572 0.6585 0.3971
0.8651 30.77 400 0.7331 0.6585 0.3971
0.8051 31.54 410 0.7182 0.6585 0.3971
0.7774 32.31 420 0.7072 0.6585 0.3971
0.749 33.08 430 0.6787 0.6585 0.3971
0.7762 33.85 440 0.6687 0.6585 0.3971
0.7223 34.62 450 0.6656 0.7480 0.6544
0.7363 35.38 460 0.6619 0.7534 0.6963
0.7039 36.15 470 0.6473 0.7371 0.6867
0.6923 36.92 480 0.6377 0.7453 0.6854
0.6667 37.69 490 0.6405 0.7317 0.6786
0.6419 38.46 500 0.6479 0.7127 0.6794
0.6511 39.23 510 0.6336 0.7344 0.6757
0.6638 40.0 520 0.6244 0.7236 0.6927
0.67 40.77 530 0.6241 0.7290 0.6795
0.616 41.54 540 0.6353 0.7182 0.6789
0.6592 42.31 550 0.6277 0.7344 0.6890
0.6146 43.08 560 0.6352 0.7236 0.6890
0.6103 43.85 570 0.6382 0.7100 0.6629
0.6099 44.62 580 0.6373 0.7100 0.6629
0.5724 45.38 590 0.6358 0.7182 0.6667
0.6134 46.15 600 0.6410 0.7073 0.6680
0.6084 46.92 610 0.6441 0.7127 0.6755
0.656 47.69 620 0.6400 0.7127 0.6727
0.6359 48.46 630 0.6405 0.7100 0.6689
0.5832 49.23 640 0.6407 0.7073 0.6621
0.5822 50.0 650 0.6403 0.7100 0.6596

Framework versions

  • Transformers 4.33.0
  • Pytorch 2.0.0
  • Datasets 2.1.0
  • Tokenizers 0.13.3
Downloads last month
7
Inference API
or
This model can be loaded on Inference API (serverless).

Finetuned from