<!-- 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. -->
torgo_xlsr_finetune-F04-2
This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the Torgo dataset. With language model added. It achieves the following results on the evaluation set:
- Loss: 0.3123
- Wer: 0.5055
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: 0.0001
- 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_steps: 1000
- num_epochs: 30
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
23.0192 | 0.89 | 500 | 3.3685 | 1.0 |
3.3848 | 1.79 | 1000 | 3.1230 | 1.0 |
2.8977 | 2.68 | 1500 | 2.7814 | 1.0 |
2.6545 | 3.57 | 2000 | 2.4289 | 1.0 |
2.0706 | 4.46 | 2500 | 1.3831 | 1.2573 |
1.4296 | 5.36 | 3000 | 0.8670 | 1.1277 |
1.0995 | 6.25 | 3500 | 0.6462 | 1.0274 |
0.8746 | 7.14 | 4000 | 0.5206 | 0.9240 |
0.7191 | 8.04 | 4500 | 0.3925 | 0.8352 |
0.6227 | 8.93 | 5000 | 0.4340 | 0.8741 |
0.5577 | 9.82 | 5500 | 0.3835 | 0.7524 |
0.4712 | 10.71 | 6000 | 0.3221 | 0.7208 |
0.4373 | 11.61 | 6500 | 0.3851 | 0.7062 |
0.3994 | 12.5 | 7000 | 0.3124 | 0.6843 |
0.3614 | 13.39 | 7500 | 0.2716 | 0.6509 |
0.3278 | 14.29 | 8000 | 0.3251 | 0.6210 |
0.3056 | 15.18 | 8500 | 0.2820 | 0.6144 |
0.3043 | 16.07 | 9000 | 0.2554 | 0.5876 |
0.2831 | 16.96 | 9500 | 0.2745 | 0.6022 |
0.2485 | 17.86 | 10000 | 0.2667 | 0.5675 |
0.2536 | 18.75 | 10500 | 0.2789 | 0.5620 |
0.2134 | 19.64 | 11000 | 0.3169 | 0.5487 |
0.2095 | 20.54 | 11500 | 0.3009 | 0.5499 |
0.1958 | 21.43 | 12000 | 0.3615 | 0.5499 |
0.2035 | 22.32 | 12500 | 0.3107 | 0.5328 |
0.1849 | 23.21 | 13000 | 0.2891 | 0.5225 |
0.191 | 24.11 | 13500 | 0.2867 | 0.5189 |
0.1505 | 25.0 | 14000 | 0.2713 | 0.5170 |
0.1583 | 25.89 | 14500 | 0.3168 | 0.5170 |
0.1607 | 26.79 | 15000 | 0.2835 | 0.5097 |
0.1438 | 27.68 | 15500 | 0.2848 | 0.5109 |
0.1446 | 28.57 | 16000 | 0.3030 | 0.5079 |
0.1362 | 29.46 | 16500 | 0.3123 | 0.5055 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1+cu116
- Datasets 1.18.3
- Tokenizers 0.13.2