<!-- 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. -->
xlsr-syntesized-turkish-2-hour
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.2104
- Wer: 0.1760
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: 0.0005
- train_batch_size: 2
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 200
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
4.7259 | 1.04 | 100 | 3.0458 | 1.0 |
2.5603 | 2.08 | 200 | 0.7478 | 0.7963 |
0.5767 | 3.12 | 300 | 0.3021 | 0.3884 |
0.3653 | 4.17 | 400 | 0.2464 | 0.3264 |
0.2777 | 5.21 | 500 | 0.2136 | 0.2785 |
0.2225 | 6.25 | 600 | 0.2319 | 0.2564 |
0.1814 | 7.29 | 700 | 0.1915 | 0.2625 |
0.1406 | 8.33 | 800 | 0.1783 | 0.2381 |
0.1242 | 9.38 | 900 | 0.1914 | 0.2227 |
0.1142 | 10.42 | 1000 | 0.2006 | 0.2308 |
0.096 | 11.46 | 1100 | 0.1945 | 0.2067 |
0.0852 | 12.5 | 1200 | 0.2025 | 0.1999 |
0.0726 | 13.54 | 1300 | 0.2206 | 0.1897 |
0.0648 | 14.58 | 1400 | 0.2083 | 0.1879 |
0.0576 | 15.62 | 1500 | 0.2081 | 0.1832 |
0.0545 | 16.67 | 1600 | 0.2157 | 0.1809 |
0.0534 | 17.71 | 1700 | 0.2135 | 0.1765 |
0.0473 | 18.75 | 1800 | 0.2111 | 0.1804 |
0.0433 | 19.79 | 1900 | 0.2104 | 0.1760 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.10.0+cu113
- Datasets 2.9.0
- Tokenizers 0.13.2