wav2vec2-korean-v3

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

wav2vec2-korean-v3

This model is a fine-tuned version of Hyuk/wav2vec2-korean-v3 on the None dataset. It achieves the following results on the evaluation set:

• eval_loss: 0.9860
• eval_wer: 0.3797
• eval_runtime: 56.5559
• eval_samples_per_second: 10.167
• eval_steps_per_second: 1.273
• epoch: 9.73
• step: 12300

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: 4
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 400
• num_epochs: 10
• mixed_precision_training: Native AMP

Framework versions

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