<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
wav2vec2-xls-r-300m-arabic_speech_commands_10s_one_speaker_40_classes
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.8103
- Accuracy: 0.7671
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: 0.0003
- 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: 80
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 3.6878 | 1.0 | 25 | 3.6888 | 0.03 |
| 3.6867 | 2.0 | 50 | 3.6867 | 0.025 |
| 3.6652 | 3.0 | 75 | 3.6260 | 0.1558 |
| 3.3912 | 4.0 | 100 | 3.3953 | 0.1254 |
| 3.103 | 5.0 | 125 | 3.1497 | 0.125 |
| 2.661 | 6.0 | 150 | 2.7077 | 0.2196 |
| 2.374 | 7.0 | 175 | 2.7057 | 0.2112 |
| 1.9198 | 8.0 | 200 | 2.5269 | 0.2929 |
| 1.7809 | 9.0 | 225 | 2.8878 | 0.2204 |
| 1.4742 | 10.0 | 250 | 2.1809 | 0.3917 |
| 1.3373 | 11.0 | 275 | 1.9832 | 0.4412 |
| 1.0601 | 12.0 | 300 | 2.0539 | 0.4958 |
| 0.9018 | 13.0 | 325 | 2.2291 | 0.46 |
| 0.6925 | 14.0 | 350 | 1.5878 | 0.5946 |
| 0.6181 | 15.0 | 375 | 2.1394 | 0.5283 |
| 0.4066 | 16.0 | 400 | 2.2009 | 0.5363 |
| 0.4346 | 17.0 | 425 | 1.9644 | 0.5625 |
| 0.3882 | 18.0 | 450 | 1.3859 | 0.6658 |
| 0.3382 | 19.0 | 475 | 1.6092 | 0.6771 |
| 0.3172 | 20.0 | 500 | 1.7496 | 0.6571 |
| 0.322 | 21.0 | 525 | 1.6505 | 0.6621 |
| 0.1848 | 22.0 | 550 | 2.1235 | 0.5933 |
| 0.2695 | 23.0 | 575 | 2.1248 | 0.6054 |
| 0.2091 | 24.0 | 600 | 2.0269 | 0.6312 |
| 0.172 | 25.0 | 625 | 1.5532 | 0.7167 |
| 0.2043 | 26.0 | 650 | 1.9791 | 0.6358 |
| 0.1744 | 27.0 | 675 | 1.4877 | 0.7458 |
| 0.1837 | 28.0 | 700 | 1.8348 | 0.6896 |
| 0.2209 | 29.0 | 725 | 2.1801 | 0.6267 |
| 0.144 | 30.0 | 750 | 1.9425 | 0.6692 |
| 0.0513 | 31.0 | 775 | 1.6531 | 0.7096 |
| 0.0494 | 32.0 | 800 | 1.8506 | 0.715 |
| 0.0697 | 33.0 | 825 | 1.9599 | 0.6933 |
| 0.1528 | 34.0 | 850 | 2.0854 | 0.6521 |
| 0.0769 | 35.0 | 875 | 2.6593 | 0.6483 |
| 0.0691 | 36.0 | 900 | 1.9098 | 0.7321 |
| 0.0401 | 37.0 | 925 | 2.0541 | 0.6967 |
| 0.0287 | 38.0 | 950 | 2.3037 | 0.6904 |
| 0.1034 | 39.0 | 975 | 1.6426 | 0.7304 |
| 0.0876 | 40.0 | 1000 | 2.1685 | 0.6775 |
| 0.0557 | 41.0 | 1025 | 2.2643 | 0.6821 |
| 0.0395 | 42.0 | 1050 | 2.0308 | 0.6979 |
| 0.1046 | 43.0 | 1075 | 2.0277 | 0.7021 |
| 0.0768 | 44.0 | 1100 | 1.7130 | 0.7371 |
| 0.048 | 45.0 | 1125 | 1.9549 | 0.7192 |
| 0.0835 | 46.0 | 1150 | 1.9024 | 0.7179 |
| 0.0505 | 47.0 | 1175 | 2.0993 | 0.7125 |
| 0.0515 | 48.0 | 1200 | 1.9806 | 0.7183 |
| 0.0556 | 49.0 | 1225 | 1.8291 | 0.7321 |
| 0.0886 | 50.0 | 1250 | 2.1479 | 0.6992 |
| 0.0769 | 51.0 | 1275 | 2.0540 | 0.7146 |
| 0.0092 | 52.0 | 1300 | 1.8446 | 0.7462 |
| 0.0032 | 53.0 | 1325 | 2.0847 | 0.7125 |
| 0.0593 | 54.0 | 1350 | 1.9553 | 0.7304 |
| 0.0053 | 55.0 | 1375 | 1.8164 | 0.74 |
| 0.0101 | 56.0 | 1400 | 1.7514 | 0.7421 |
| 0.0155 | 57.0 | 1425 | 1.6395 | 0.7604 |
| 0.0035 | 58.0 | 1450 | 1.7393 | 0.7504 |
| 0.0019 | 59.0 | 1475 | 1.8103 | 0.7671 |
| 0.0144 | 60.0 | 1500 | 1.8234 | 0.7588 |
| 0.0028 | 61.0 | 1525 | 1.8479 | 0.7529 |
| 0.0306 | 62.0 | 1550 | 1.7948 | 0.7454 |
| 0.0028 | 63.0 | 1575 | 1.7417 | 0.7562 |
| 0.0095 | 64.0 | 1600 | 1.6973 | 0.7592 |
| 0.0086 | 65.0 | 1625 | 1.9997 | 0.7342 |
| 0.0953 | 66.0 | 1650 | 1.8202 | 0.7538 |
| 0.0018 | 67.0 | 1675 | 1.8316 | 0.7533 |
| 0.0053 | 68.0 | 1700 | 1.8916 | 0.7475 |
| 0.004 | 69.0 | 1725 | 1.8794 | 0.7521 |
| 0.0169 | 70.0 | 1750 | 1.8215 | 0.7533 |
| 0.0013 | 71.0 | 1775 | 1.7565 | 0.7508 |
| 0.0008 | 72.0 | 1800 | 1.8171 | 0.7454 |
| 0.0011 | 73.0 | 1825 | 1.8354 | 0.7479 |
| 0.0025 | 74.0 | 1850 | 1.8283 | 0.7488 |
| 0.0013 | 75.0 | 1875 | 1.8876 | 0.7412 |
| 0.0415 | 76.0 | 1900 | 1.8789 | 0.7454 |
| 0.0341 | 77.0 | 1925 | 1.8665 | 0.7512 |
| 0.0149 | 78.0 | 1950 | 1.8579 | 0.7488 |
| 0.0018 | 79.0 | 1975 | 1.8571 | 0.7488 |
| 0.008 | 80.0 | 2000 | 1.8596 | 0.7479 |
Framework versions
- Transformers 4.21.1
- Pytorch 1.12.1
- Datasets 2.4.0
- Tokenizers 0.12.1
