metadata
base_model: facebook/mms-1b-all
datasets:
- common_voice_17_0
library_name: transformers
license: cc-by-nc-4.0
metrics:
- wer
- bleu
tags:
- generated_from_trainer
model-index:
- name: wav2vec2-mms-1b-CV17.0-training_set_variations
results:
- task:
type: automatic-speech-recognition
name: Automatic Speech Recognition
dataset:
name: common_voice_17_0
type: common_voice_17_0
config: ta
split: validation
args: ta
metrics:
- type: wer
value: 0.5038262932638631
name: Wer
- type: bleu
value: 0.25352723931305554
name: Bleu
wav2vec2-mms-1b-CV17.0-training_set_variations
This model is a fine-tuned version of facebook/mms-1b-all on the common_voice_17_0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.5467
- Wer: 0.5038
- Cer: 0.0871
- Bleu: 0.2535
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: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.15
- training_steps: 2000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer | Bleu |
|---|---|---|---|---|---|---|
| 7.5761 | 100.0 | 100 | 0.3128 | 0.4455 | 0.0729 | 0.3095 |
| 0.067 | 200.0 | 200 | 0.3441 | 0.4234 | 0.0706 | 0.3389 |
| 0.0307 | 300.0 | 300 | 0.3906 | 0.4489 | 0.0749 | 0.3100 |
| 0.0251 | 400.0 | 400 | 0.4461 | 0.4745 | 0.0802 | 0.2744 |
| 0.0205 | 500.0 | 500 | 0.4579 | 0.4834 | 0.0820 | 0.2714 |
| 0.0166 | 600.0 | 600 | 0.4550 | 0.4742 | 0.0823 | 0.2837 |
| 0.0123 | 700.0 | 700 | 0.5467 | 0.5038 | 0.0871 | 0.2535 |
Framework versions
- Transformers 4.44.2
- Pytorch 2.4.1+cu121
- Datasets 3.0.0
- Tokenizers 0.19.1