Ussen's picture
End of training
46c0ac7
metadata
license: cc-by-nc-4.0
base_model: facebook/mms-1b-all
tags:
  - generated_from_trainer
datasets:
  - audiofolder
metrics:
  - wer
model-index:
  - name: wav2vec2-mms-1b-all-swc-kat6
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: audiofolder
          type: audiofolder
          config: default
          split: train
          args: default
        metrics:
          - name: Wer
            type: wer
            value: 0.37098445595854923

wav2vec2-mms-1b-all-swc-kat6

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

  • Loss: inf
  • Wer: 0.3710

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: 1
  • 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: 400
  • num_epochs: 4

Training results

Training Loss Epoch Step Validation Loss Wer
2.8349 0.07 400 inf 0.4899
0.8039 0.15 800 inf 0.4912
0.7335 0.22 1200 inf 0.4482
0.8395 0.3 1600 inf 0.4829
0.7626 0.37 2000 inf 0.4635
0.9035 0.44 2400 inf 0.4391
0.6281 0.52 2800 inf 0.4668
0.6756 0.59 3200 inf 0.4254
0.7866 0.67 3600 inf 0.4168
0.7413 0.74 4000 inf 0.4168
0.749 0.81 4400 inf 0.4148
0.8165 0.89 4800 inf 0.4241
0.7302 0.96 5200 inf 0.4124
0.7376 1.04 5600 inf 0.4368
0.6833 1.11 6000 inf 0.3953
0.6463 1.19 6400 inf 0.4363
0.7236 1.26 6800 inf 0.4383
0.8837 1.33 7200 inf 0.4811
0.6854 1.41 7600 inf 0.3930
0.6985 1.48 8000 inf 0.3979
0.7139 1.56 8400 inf 0.3977
0.6338 1.63 8800 inf 0.4039
0.7227 1.7 9200 inf 0.3922
0.6843 1.78 9600 inf 0.4111
0.6948 1.85 10000 inf 0.4093
0.6867 1.93 10400 inf 0.3927
0.5753 2.0 10800 inf 0.4080
0.6865 2.07 11200 inf 0.3938
0.6155 2.15 11600 inf 0.3920
0.6743 2.22 12000 inf 0.3977
0.5801 2.3 12400 inf 0.3801
0.8216 2.37 12800 inf 0.3917
0.6199 2.44 13200 inf 0.4052
0.6268 2.52 13600 inf 0.3811
0.6505 2.59 14000 inf 0.3855
0.6578 2.67 14400 inf 0.3933
0.6442 2.74 14800 inf 0.3868
0.5904 2.81 15200 inf 0.3782
0.6249 2.89 15600 inf 0.3788
0.5879 2.96 16000 inf 0.3904
0.4844 3.04 16400 inf 0.3728
0.6309 3.11 16800 inf 0.3687
0.5825 3.19 17200 inf 0.3663
0.7171 3.26 17600 inf 0.3772
0.5471 3.33 18000 inf 0.3718
0.5029 3.41 18400 inf 0.3756
0.5605 3.48 18800 inf 0.3751
0.5582 3.56 19200 inf 0.3728
0.6358 3.63 19600 inf 0.3712
0.4977 3.7 20000 inf 0.3655
0.4828 3.78 20400 inf 0.3671
0.6554 3.85 20800 inf 0.3689
0.561 3.93 21200 inf 0.3702
0.5515 4.0 21600 inf 0.3710

Framework versions

  • Transformers 4.33.0.dev0
  • Pytorch 1.13.1+cu117
  • Datasets 2.13.1
  • Tokenizers 0.12.1