metadata

license: apache-2.0
language: bas
tags:
  - automatic-speech-recognition
  - common_voice
  - generated_from_trainer
  - bas
  - robust-speech-event
  - hf-asr-leaderboard
datasets:
  - mozilla-foundation/common_voice_8_0
model-index:
  - name: wav2vec2-xls-r-300m-bas-CV8-v2
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Common Voice 8
          type: mozilla-foundation/common_voice_8_0
          args: bas
        metrics:
          - name: Test WER
            type: wer
            value: 56.97

wav2vec2-xls-r-300m-bas-CV8-v2

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: 0.6121
Wer: 0.5697

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: 0.0001
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: 300
num_epochs: 90
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
6.5211	16.13	500	1.2661	0.9153
0.7026	32.25	1000	0.6245	0.6516
0.3752	48.38	1500	0.6039	0.6148
0.2752	64.51	2000	0.6080	0.5808
0.2155	80.63	2500	0.6121	0.5697

Framework versions

Transformers 4.11.3
Pytorch 1.10.0+cu111
Datasets 1.18.3
Tokenizers 0.10.3