xls-r-300m-finetune-vi-en-200hrs-lr2e-05-epochs20-char-based-non_freeze

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

• Loss: nan
• Wer: 1.0

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: 2e-05
• train_batch_size: 10
• eval_batch_size: 4
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 40
• total_eval_batch_size: 16
• optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
3.5491	1.48	2500	3.7343	1.0
3.6495	2.97	5000	3.6478	1.0
3.6513	4.45	7500	3.6478	1.0
3.6419	5.93	10000	3.6478	1.0
3.656	7.42	12500	3.6478	1.0
0.0	8.9	15000	nan	1.0
0.0	10.39	17500	nan	1.0
0.0	11.87	20000	nan	1.0
0.0	13.35	22500	nan	1.0
0.0	14.84	25000	nan	1.0
0.0	16.32	27500	nan	1.0
0.0	17.8	30000	nan	1.0
0.0	19.29	32500	nan	1.0

Framework versions

• Transformers 4.34.0
• Pytorch 2.0.1
• Datasets 2.14.5
• Tokenizers 0.14.1