|
--- |
|
library_name: transformers |
|
license: cc-by-nc-4.0 |
|
base_model: facebook/mms-1b-all |
|
tags: |
|
- automatic-speech-recognition |
|
- genbed |
|
- mms |
|
- generated_from_trainer |
|
metrics: |
|
- wer |
|
model-index: |
|
- name: mms-1b-bemgen-combined-model |
|
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. --> |
|
|
|
# mms-1b-bemgen-combined-model |
|
|
|
This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on the GENBED - BEM dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.2591 |
|
- Wer: 0.4134 |
|
|
|
## 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 |
|
- 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: 30.0 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | |
|
|:-------------:|:------:|:----:|:---------------:|:------:| |
|
| 6.7553 | 0.0516 | 100 | 0.8774 | 0.8476 | |
|
| 0.5648 | 0.1031 | 200 | 0.3408 | 0.5032 | |
|
| 0.4827 | 0.1547 | 300 | 0.3261 | 0.4930 | |
|
| 0.4321 | 0.2063 | 400 | 0.3036 | 0.4854 | |
|
| 0.4168 | 0.2579 | 500 | 0.2989 | 0.4783 | |
|
| 0.3965 | 0.3094 | 600 | 0.2907 | 0.4513 | |
|
| 0.4199 | 0.3610 | 700 | 0.2926 | 0.4718 | |
|
| 0.3975 | 0.4126 | 800 | 0.2886 | 0.4459 | |
|
| 0.3839 | 0.4642 | 900 | 0.2908 | 0.4722 | |
|
| 0.3673 | 0.5157 | 1000 | 0.2836 | 0.4445 | |
|
| 0.3777 | 0.5673 | 1100 | 0.2784 | 0.4365 | |
|
| 0.3764 | 0.6189 | 1200 | 0.2791 | 0.4278 | |
|
| 0.3918 | 0.6704 | 1300 | 0.2757 | 0.4251 | |
|
| 0.3669 | 0.7220 | 1400 | 0.2721 | 0.4182 | |
|
| 0.377 | 0.7736 | 1500 | 0.2728 | 0.4757 | |
|
| 0.4174 | 0.8252 | 1600 | 0.2684 | 0.4242 | |
|
| 0.3641 | 0.8767 | 1700 | 0.2649 | 0.4195 | |
|
| 0.3882 | 0.9283 | 1800 | 0.2647 | 0.4125 | |
|
| 0.3861 | 0.9799 | 1900 | 0.2668 | 0.4425 | |
|
| 0.3647 | 1.0315 | 2000 | 0.2675 | 0.4246 | |
|
| 0.3467 | 1.0830 | 2100 | 0.2629 | 0.4098 | |
|
| 0.3579 | 1.1346 | 2200 | 0.2587 | 0.4186 | |
|
| 0.3544 | 1.1862 | 2300 | 0.2609 | 0.4127 | |
|
| 0.35 | 1.2378 | 2400 | 0.2592 | 0.4062 | |
|
| 0.3519 | 1.2893 | 2500 | 0.2591 | 0.4135 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.48.0.dev0 |
|
- Pytorch 2.5.1+cu124 |
|
- Datasets 3.2.0 |
|
- Tokenizers 0.21.0 |
|
|
