<!-- 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_5_classes_TTS
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: 0.0035
- Accuracy: 1.0
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 |
---|---|---|---|---|
1.6044 | 0.99 | 34 | 1.6081 | 0.2 |
1.4831 | 1.99 | 68 | 1.1160 | 0.5667 |
0.6296 | 2.99 | 102 | 0.5423 | 0.7767 |
0.1944 | 3.99 | 136 | 0.1192 | 0.9633 |
0.1273 | 4.99 | 170 | 0.9873 | 0.7533 |
0.0207 | 5.99 | 204 | 0.2619 | 0.9467 |
0.0603 | 6.99 | 238 | 0.0470 | 0.9867 |
0.1368 | 7.99 | 272 | 0.3046 | 0.9367 |
0.0323 | 8.99 | 306 | 0.2696 | 0.9367 |
0.0562 | 9.99 | 340 | 0.0420 | 0.9867 |
0.0739 | 10.99 | 374 | 0.0035 | 1.0 |
0.0169 | 11.99 | 408 | 1.1230 | 0.8367 |
0.2708 | 12.99 | 442 | 1.8698 | 0.7167 |
0.1253 | 13.99 | 476 | 0.2689 | 0.93 |
0.0856 | 14.99 | 510 | 0.4139 | 0.92 |
0.0512 | 15.99 | 544 | 0.2122 | 0.9567 |
0.057 | 16.99 | 578 | 0.2451 | 0.94 |
0.1541 | 17.99 | 612 | 0.3384 | 0.93 |
0.0032 | 18.99 | 646 | 0.0901 | 0.9767 |
0.0017 | 19.99 | 680 | 0.2857 | 0.9267 |
0.049 | 20.99 | 714 | 0.3468 | 0.93 |
0.0316 | 21.99 | 748 | 0.0818 | 0.9733 |
0.0112 | 22.99 | 782 | 0.2826 | 0.9433 |
0.0012 | 23.99 | 816 | 0.0419 | 0.9933 |
0.011 | 24.99 | 850 | 0.2193 | 0.95 |
0.0033 | 25.99 | 884 | 0.2857 | 0.9467 |
0.0345 | 26.99 | 918 | 0.3890 | 0.9367 |
0.0432 | 27.99 | 952 | 0.3312 | 0.93 |
0.0023 | 28.99 | 986 | 0.5031 | 0.9133 |
0.0005 | 29.99 | 1020 | 0.4338 | 0.92 |
0.0404 | 30.99 | 1054 | 0.2566 | 0.9533 |
0.038 | 31.99 | 1088 | 0.3244 | 0.95 |
0.0009 | 32.99 | 1122 | 0.0655 | 0.9833 |
0.0012 | 33.99 | 1156 | 0.0503 | 0.9867 |
0.0004 | 34.99 | 1190 | 0.8135 | 0.8933 |
0.0006 | 35.99 | 1224 | 0.3515 | 0.9333 |
0.0005 | 36.99 | 1258 | 0.2831 | 0.96 |
0.0004 | 37.99 | 1292 | 0.3297 | 0.95 |
0.0005 | 38.99 | 1326 | 0.2892 | 0.9533 |
0.0059 | 39.99 | 1360 | 0.2430 | 0.96 |
0.0506 | 40.99 | 1394 | 0.3800 | 0.94 |
0.0004 | 41.99 | 1428 | 0.3654 | 0.9467 |
0.043 | 42.99 | 1462 | 0.2559 | 0.9567 |
0.0456 | 43.99 | 1496 | 0.1877 | 0.9633 |
0.0004 | 44.99 | 1530 | 0.1706 | 0.9667 |
0.0461 | 45.99 | 1564 | 0.0958 | 0.9867 |
0.0021 | 46.99 | 1598 | 0.1074 | 0.9833 |
0.0006 | 47.99 | 1632 | 0.1116 | 0.9833 |
0.0002 | 48.99 | 1666 | 0.1106 | 0.9867 |
0.0002 | 49.99 | 1700 | 0.1113 | 0.9867 |
0.0005 | 50.99 | 1734 | 0.1115 | 0.9867 |
0.0002 | 51.99 | 1768 | 0.1122 | 0.9867 |
0.001 | 52.99 | 1802 | 0.1389 | 0.98 |
0.0002 | 53.99 | 1836 | 0.1306 | 0.9833 |
0.0002 | 54.99 | 1870 | 0.1069 | 0.9833 |
0.0002 | 55.99 | 1904 | 0.1053 | 0.9833 |
0.0004 | 56.99 | 1938 | 0.1872 | 0.9767 |
0.0002 | 57.99 | 1972 | 0.2230 | 0.9633 |
0.0075 | 58.99 | 2006 | 0.1857 | 0.9733 |
0.0001 | 59.99 | 2040 | 0.1630 | 0.9767 |
0.0002 | 60.99 | 2074 | 0.1424 | 0.9833 |
0.0616 | 61.99 | 2108 | 0.1600 | 0.9767 |
0.0002 | 62.99 | 2142 | 0.1554 | 0.98 |
0.0004 | 63.99 | 2176 | 0.2557 | 0.9667 |
0.0002 | 64.99 | 2210 | 0.2397 | 0.9667 |
0.0002 | 65.99 | 2244 | 0.3841 | 0.94 |
0.0002 | 66.99 | 2278 | 0.4385 | 0.94 |
0.0001 | 67.99 | 2312 | 0.4248 | 0.94 |
0.0001 | 68.99 | 2346 | 0.4203 | 0.94 |
0.0002 | 69.99 | 2380 | 0.3655 | 0.9433 |
0.0001 | 70.99 | 2414 | 0.3608 | 0.9433 |
0.0001 | 71.99 | 2448 | 0.3602 | 0.9433 |
0.0073 | 72.99 | 2482 | 0.2625 | 0.96 |
0.0002 | 73.99 | 2516 | 0.2479 | 0.9633 |
0.0001 | 74.99 | 2550 | 0.2474 | 0.9667 |
0.0002 | 75.99 | 2584 | 0.2474 | 0.9667 |
0.0001 | 76.99 | 2618 | 0.2464 | 0.9667 |
0.003 | 77.99 | 2652 | 0.2439 | 0.9667 |
0.0002 | 78.99 | 2686 | 0.2433 | 0.9667 |
0.0002 | 79.99 | 2720 | 0.2433 | 0.9667 |
Framework versions
- Transformers 4.21.1
- Pytorch 1.12.1
- Datasets 2.4.0
- Tokenizers 0.12.1