wav2vec2-large-xls-r-300m-gn-pt

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice_13_0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.6576
• Wer: 0.5965

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: 100
• num_epochs: 35

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
4.7255	4.85	400	3.0288	1.0
1.1814	9.7	800	0.5687	0.7317
0.3099	14.55	1200	0.6297	0.6828
0.1719	19.39	1600	0.7157	0.6992
0.1185	24.24	2000	0.6896	0.6537
0.0871	29.09	2400	0.7071	0.6215
0.0647	33.94	2800	0.6576	0.5965

Framework versions

• Transformers 4.34.1
• Pytorch 2.0.1+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1