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

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: 0.4411
• Wer: 0.3271

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
3.8286	3.67	400	0.6899	0.7462
0.4378	7.34	800	0.4803	0.5127
0.2073	11.01	1200	0.4640	0.4584
0.1386	14.68	1600	0.4355	0.4252
0.1058	18.35	2000	0.4476	0.3789
0.0819	22.02	2400	0.4248	0.3543
0.0666	25.69	2800	0.4276	0.3399
0.0525	29.36	3200	0.4411	0.3271

Framework versions

• Transformers 4.11.3
• Pytorch 1.10.0+cu113
• Datasets 1.18.3
• Tokenizers 0.10.3