This checkpoint is a model fine-tuned for multi-lingual ASR using Facebook's Massive Multilingual Speech project. This checkpoint is based on the Wav2Vec2 architecture and makes use of adapter models to transcribe 1000+ languages. The checkpoint consists of 1 billion parameters and has been fine-tuned from facebook/mms-1b on Yoruba.

Intended uses & limitations

More information needed

Training and evaluation data

Common Voice 16.1 Yoruba data

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
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
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
4.8923	0.77	100	0.7710	0.7413
0.7507	1.54	200	0.7249	0.7585
0.7033	2.31	300	0.7105	0.7247
0.6888	3.08	400	0.6829	0.7229
0.6471	3.85	500	0.6682	0.6802

Framework versions

Transformers 4.38.0.dev0
Pytorch 2.1.0+cu121
Datasets 2.16.1
Tokenizers 0.15.0

ogbi/wav2vec2-large-mms-1b-yoruba-test

wav2vec2-large-mms-1b-yoruba-test

Model description