README.md

metadata

license: apache-2.0
base_model: facebook/wav2vec2-large-xlsr-53
tags:
  - generated_from_trainer
datasets:
  - zeroth_korean
metrics:
  - wer
model-index:
  - name: wav2vec2-large-xlsr-test
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: zeroth_korean
          type: zeroth_korean
          config: clean
          split: None
          args: clean
        metrics:
          - name: Wer
            type: wer
            value: 1

wav2vec2-large-xlsr-test

This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the zeroth_korean dataset. It achieves the following results on the evaluation set:

• Loss: 3.8883
• 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: 0.0003
• train_batch_size: 64
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1000
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
4.3921	50.0	1000	3.8883	1.0

Framework versions

• Transformers 4.40.2
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1