|
--- |
|
library_name: transformers |
|
license: cc-by-nc-4.0 |
|
base_model: facebook/mms-1b-all |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- wer |
|
model-index: |
|
- name: wav2vec2-large-mms-1b-dzo_M2 |
|
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. --> |
|
|
|
# wav2vec2-large-mms-1b-dzo_M2 |
|
|
|
This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.4519 |
|
- Wer: 0.4212 |
|
|
|
## 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: 0.005 |
|
- train_batch_size: 8 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- optimizer: Use OptimizerNames.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: 500 |
|
- num_epochs: 8 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | |
|
|:-------------:|:-----:|:----:|:---------------:|:------:| |
|
| 8.852 | 1.0 | 25 | 5.4081 | 1.0355 | |
|
| 3.2709 | 2.0 | 50 | 2.9079 | 0.9976 | |
|
| 2.0551 | 3.0 | 75 | 1.3991 | 0.7921 | |
|
| 1.2379 | 4.0 | 100 | 0.9750 | 0.6851 | |
|
| 0.9935 | 5.0 | 125 | 0.7704 | 0.5634 | |
|
| 0.8459 | 6.0 | 150 | 0.6053 | 0.4764 | |
|
| 0.7386 | 7.0 | 175 | 0.5166 | 0.4373 | |
|
| 0.6474 | 8.0 | 200 | 0.4519 | 0.4212 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.49.0.dev0 |
|
- Pytorch 2.5.1+cu121 |
|
- Datasets 3.2.0 |
|
- Tokenizers 0.21.0 |
|
|
