metadata
license: cc-by-nc-4.0
base_model: facebook/mms-1b-all
tags:
- generated_from_trainer
datasets:
- audiofolder
metrics:
- wer
model-index:
- name: wav2vec2-large-mms-1b-all-lingala-ojpl-4
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: audiofolder
type: audiofolder
config: default
split: train
args: default
metrics:
- name: Wer
type: wer
value: 0.2868821999256782
wav2vec2-large-mms-1b-all-lingala-ojpl-4
This model is a fine-tuned version of facebook/mms-1b-all on the audiofolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.7849
- Wer: 0.2869
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: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 4
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 3.5152 | 0.54 | 100 | 0.9315 | 0.3430 |
| 0.4837 | 1.08 | 200 | 0.9377 | 0.3133 |
| 0.5943 | 1.61 | 300 | 0.8318 | 0.3177 |
| 0.6232 | 2.15 | 400 | 0.7491 | 0.3107 |
| 0.5186 | 2.69 | 500 | 0.7888 | 0.2899 |
| 0.6024 | 3.23 | 600 | 0.7674 | 0.2913 |
| 0.4522 | 3.76 | 700 | 0.7849 | 0.2869 |
Framework versions
- Transformers 4.32.0.dev0
- Pytorch 1.13.1+cu117
- Datasets 2.13.1
- Tokenizers 0.13.3