metadata
language:
- skr
license: cc-by-nc-4.0
base_model: facebook/mms-300m
tags:
- skr
- generated_from_trainer
datasets:
- mozilla-foundation/common_voice_11_0
metrics:
- wer
model-index:
- name: Saraiki Speech
results: []
Saraiki Speech
This model is a fine-tuned version of facebook/mms-300m on the common_voice dataset. It achieves the following results on the evaluation set:
- Loss: 3.4467
- Wer: 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: 0.001
- 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: 100
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 3.5576 | 0.58 | 100 | 3.5468 | 1.0 |
| 3.2271 | 1.16 | 200 | 3.6465 | 1.0 |
| 3.3492 | 1.73 | 300 | 3.8570 | 1.0 |
| 3.4442 | 2.31 | 400 | 3.7065 | 1.0 |
| 3.29 | 2.89 | 500 | 3.5289 | 1.0 |
| 3.2951 | 3.47 | 600 | 3.7043 | 1.0 |
| 3.2919 | 4.05 | 700 | 3.6748 | 1.0 |
| 3.301 | 4.62 | 800 | 3.4422 | 1.0 |
| 3.2103 | 5.2 | 900 | 3.4955 | 1.0 |
| 3.2728 | 5.78 | 1000 | 3.6059 | 1.0 |
| 3.2458 | 6.36 | 1100 | 3.4087 | 1.0 |
| 3.244 | 6.94 | 1200 | 3.4352 | 1.0 |
| 3.2562 | 7.51 | 1300 | 3.4648 | 1.0 |
| 3.2116 | 8.09 | 1400 | 3.4618 | 1.0 |
| 3.2268 | 8.67 | 1500 | 3.4313 | 1.0 |
| 3.2387 | 9.25 | 1600 | 3.4246 | 1.0 |
| 3.1921 | 9.83 | 1700 | 3.4467 | 1.0 |
Framework versions
- Transformers 4.36.0.dev0
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0