wav2vec2-large-xlsr-oria-v0

This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the fleurs dataset. It achieves the following results on the evaluation set:

• Loss: 0.6050
• Wer: 0.4972

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: 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_steps: 500
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
6.6722	2.1505	100	5.4667	1.0
3.4956	4.3011	200	3.4905	1.0
3.4526	6.4516	300	3.4624	1.0
3.4548	8.6022	400	3.4494	1.0
3.4352	10.7527	500	3.4267	1.0
3.0335	12.9032	600	2.8300	1.0
1.021	15.0538	700	0.9941	0.7938
0.6175	17.2043	800	0.7318	0.6385
0.5257	19.3548	900	0.6485	0.5820
0.4232	21.5054	1000	0.6105	0.5430
0.3202	23.6559	1100	0.5906	0.5192
0.2767	25.8065	1200	0.6025	0.5079
0.2679	27.9570	1300	0.6050	0.4972

Framework versions

• Transformers 4.45.0.dev0
• Pytorch 2.4.1+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1