README.md

metadata

license: cc-by-nc-4.0
base_model: facebook/mms-1b-l1107
tags:
  - generated_from_trainer
datasets:
  - common_voice_13_0
metrics:
  - wer
model-index:
  - name: wav2vec2-large-mms-1b-korean-colab
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: common_voice_13_0
          type: common_voice_13_0
          config: ko
          split: test
          args: ko
        metrics:
          - name: Wer
            type: wer
            value: 0.9859013091641491

wav2vec2-large-mms-1b-korean-colab

This model is a fine-tuned version of facebook/mms-1b-l1107 on the common_voice_13_0 dataset. It achieves the following results on the evaluation set:

• Loss: 7.8369
• Wer: 0.9859

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: 0.001
• train_batch_size: 8
• 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: 100
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
11.2038	2.63	100	7.7632	0.9869
5.9383	5.26	200	8.0920	0.9909
5.6379	7.89	300	7.8850	0.9930
5.0749	10.53	400	7.8490	0.9940
3.791	13.16	500	7.4713	0.9849
3.3555	15.79	600	7.6595	0.9799
2.5415	18.42	700	7.8369	0.9859

Framework versions

• Transformers 4.31.0.dev0
• Pytorch 2.0.1+cu118
• Datasets 2.13.1
• Tokenizers 0.13.3