|
|
--- |
|
|
library_name: transformers |
|
|
license: cc-by-nc-4.0 |
|
|
base_model: facebook/mms-1b-all |
|
|
tags: |
|
|
- generated_from_trainer |
|
|
metrics: |
|
|
- wer |
|
|
model-index: |
|
|
- name: Yakut-ASR |
|
|
results: [] |
|
|
--- |
|
|
|
|
|
<!-- 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. --> |
|
|
|
|
|
# Yakut-ASR |
|
|
|
|
|
This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on an unknown dataset. |
|
|
It achieves the following results on the evaluation set: |
|
|
- Loss: 0.2140 |
|
|
- Wer: 0.2772 |
|
|
|
|
|
## 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: 4 |
|
|
- eval_batch_size: 8 |
|
|
- seed: 42 |
|
|
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments |
|
|
- lr_scheduler_type: linear |
|
|
- lr_scheduler_warmup_steps: 100 |
|
|
- num_epochs: 4 |
|
|
- mixed_precision_training: Native AMP |
|
|
|
|
|
### Training results |
|
|
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | |
|
|
|:-------------:|:------:|:----:|:---------------:|:------:| |
|
|
| 6.0192 | 0.2132 | 100 | 4.4580 | 0.9999 | |
|
|
| 3.5904 | 0.4264 | 200 | 3.1478 | 1.0 | |
|
|
| 2.5173 | 0.6397 | 300 | 0.3987 | 0.4625 | |
|
|
| 0.2948 | 0.8529 | 400 | 0.2442 | 0.3075 | |
|
|
| 0.2407 | 1.0661 | 500 | 0.2367 | 0.3170 | |
|
|
| 0.2278 | 1.2793 | 600 | 0.2280 | 0.2914 | |
|
|
| 0.2279 | 1.4925 | 700 | 0.2341 | 0.2963 | |
|
|
| 0.2097 | 1.7058 | 800 | 0.2303 | 0.3138 | |
|
|
| 0.2317 | 1.9190 | 900 | 0.2253 | 0.2889 | |
|
|
| 0.1898 | 2.1322 | 1000 | 0.2187 | 0.2795 | |
|
|
| 0.1925 | 2.3454 | 1100 | 0.2262 | 0.2951 | |
|
|
| 0.211 | 2.5586 | 1200 | 0.2205 | 0.2909 | |
|
|
| 0.1942 | 2.7719 | 1300 | 0.2192 | 0.2792 | |
|
|
| 0.169 | 2.9851 | 1400 | 0.2213 | 0.2835 | |
|
|
| 0.178 | 3.1983 | 1500 | 0.2148 | 0.2795 | |
|
|
| 0.1862 | 3.4115 | 1600 | 0.2145 | 0.2803 | |
|
|
| 0.1896 | 3.6247 | 1700 | 0.2154 | 0.2788 | |
|
|
| 0.1813 | 3.8380 | 1800 | 0.2140 | 0.2772 | |
|
|
|
|
|
|
|
|
### Framework versions |
|
|
|
|
|
- Transformers 4.49.0.dev0 |
|
|
- Pytorch 2.5.1+cu121 |
|
|
- Datasets 3.2.0 |
|
|
- Tokenizers 0.21.0 |
|
|
|
