<!-- 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. -->
AIRI UZ jamoasi a'zolari tomonidan qilingan STT ning eng kichik versiyasi. Bu yaxshi versiya emas faqat 15ta epoch o'qitilgan!
Azizbek Biyimbetov, Foziljon To'lqinov, Shaxboz Zohidov, Abduraxim Jabborov, Yahyoxon Rahimov
This model is a fine-tuned version of facebook/wav2vec2-base on the MOZILLA-FOUNDATION/COMMON_VOICE_10_0 - UZ dataset. It achieves the following results on the evaluation set:
- Loss: 0.5763
- Wer: 0.4502
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: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 2
- total_train_batch_size: 8
- total_eval_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 15.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.4736 | 3.4 | 10000 | 0.6247 | 0.6362 |
| 0.3392 | 6.8 | 20000 | 0.7254 | 0.5605 |
| 0.2085 | 10.19 | 30000 | 0.5465 | 0.5097 |
| 0.1387 | 13.59 | 40000 | 0.5984 | 0.4632 |
Framework versions
- Transformers 4.22.0.dev0
- Pytorch 1.10.0+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1
