wav2vec2-large-xls-r-300m-turkish-colab-main

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

• Loss: 0.3764
• Wer: 0.3168

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

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
3.822	3.67	400	0.6508	0.6687
0.399	7.34	800	0.4276	0.4480
0.1905	11.01	1200	0.3999	0.4225
0.1249	14.68	1600	0.4302	0.3910
0.0978	18.35	2000	0.3766	0.3682
0.0773	22.02	2400	0.3877	0.3483
0.0597	25.69	2800	0.3833	0.3268
0.0467	29.36	3200	0.3764	0.3168

Framework versions

• Transformers 4.34.1
• Pytorch 2.0.1+cu118
• Datasets 2.14.5
• Tokenizers 0.14.1