metadata
license: cc-by-nc-4.0
base_model: facebook/mms-1b-all
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: mms-SADA
results: []
mms-SADA
This model is a fine-tuned version of facebook/mms-1b-all on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.1720
- Wer: 0.6627
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: 1e-05
- train_batch_size: 14
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant_with_warmup
- lr_scheduler_warmup_steps: 50
- training_steps: 10000
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.7672 | 0.0 | 250 | 1.3082 | 0.6868 |
| 1.7864 | 0.01 | 500 | 1.3002 | 0.6852 |
| 1.445 | 0.01 | 750 | 1.2947 | 0.6846 |
| 1.6083 | 0.01 | 1000 | 1.2921 | 0.6831 |
| 1.6405 | 0.02 | 1250 | 1.2864 | 0.6819 |
| 1.6112 | 0.02 | 1500 | 1.2774 | 0.6826 |
| 1.5307 | 0.02 | 1750 | 1.2730 | 0.6812 |
| 1.8135 | 0.02 | 2000 | 1.2694 | 0.6795 |
| 1.6133 | 0.03 | 2250 | 1.2660 | 0.6783 |
| 1.8358 | 0.03 | 2500 | 1.2633 | 0.6758 |
| 1.507 | 0.03 | 2750 | 1.2563 | 0.6774 |
| 1.7197 | 0.04 | 3000 | 1.2553 | 0.6750 |
| 1.5191 | 0.04 | 3250 | 1.2498 | 0.6737 |
| 1.4389 | 0.04 | 3500 | 1.2478 | 0.6734 |
| 1.6184 | 0.05 | 3750 | 1.2401 | 0.6723 |
| 1.6814 | 0.05 | 4000 | 1.2357 | 0.6716 |
| 1.4742 | 0.05 | 4250 | 1.2304 | 0.6708 |
| 1.4276 | 0.06 | 4500 | 1.2302 | 0.6700 |
| 1.4855 | 0.06 | 4750 | 1.2224 | 0.6693 |
| 1.4409 | 0.06 | 5000 | 1.2197 | 0.6693 |
| 1.4562 | 0.07 | 5250 | 1.2162 | 0.6688 |
| 1.5353 | 0.07 | 5500 | 1.2119 | 0.6689 |
| 1.5601 | 0.07 | 5750 | 1.2105 | 0.6696 |
| 1.4666 | 0.07 | 6000 | 1.2066 | 0.6679 |
| 1.6642 | 0.08 | 6250 | 1.2010 | 0.6687 |
| 1.5008 | 0.08 | 6500 | 1.2005 | 0.6669 |
| 1.6213 | 0.08 | 6750 | 1.2008 | 0.6665 |
| 1.7335 | 0.09 | 7000 | 1.1938 | 0.6675 |
| 1.421 | 0.09 | 7250 | 1.1921 | 0.6666 |
| 1.6255 | 0.09 | 7500 | 1.1919 | 0.6645 |
| 1.4785 | 0.1 | 7750 | 1.1895 | 0.6646 |
| 1.6736 | 0.1 | 8000 | 1.1918 | 0.6634 |
| 1.4629 | 0.1 | 8250 | 1.1841 | 0.6645 |
| 1.6599 | 0.11 | 8500 | 1.1832 | 0.6628 |
| 1.4726 | 0.11 | 8750 | 1.1790 | 0.6649 |
| 1.6825 | 0.11 | 9000 | 1.1774 | 0.6636 |
| 1.6216 | 0.11 | 9250 | 1.1815 | 0.6630 |
| 1.4291 | 0.12 | 9500 | 1.1768 | 0.6637 |
| 1.2947 | 0.12 | 9750 | 1.1743 | 0.6623 |
| 1.4702 | 0.12 | 10000 | 1.1720 | 0.6627 |
Framework versions
- Transformers 4.33.2
- Pytorch 2.0.1
- Datasets 2.19.1
- Tokenizers 0.13.3