metadata

license: apache-2.0
base_model: facebook/wav2vec2-large-xlsr-53
tags:
  - generated_from_trainer
datasets:
  - common_voice_6_1
metrics:
  - wer
model-index:
  - name: wav2vec2-large-morrnla-ar-colab
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: common_voice_6_1
          type: common_voice_6_1
          config: ar
          split: test[10:20]
          args: ar
        metrics:
          - name: Wer
            type: wer
            value: 1

wav2vec2-large-morrnla-ar-colab

This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the common_voice_6_1 dataset. It achieves the following results on the evaluation set:

Loss: 3.5524
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: 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: 20
num_epochs: 20
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
16.4983	3.33	10	23.6467	1.0
8.0393	6.67	20	7.4047	1.0
4.2297	10.0	30	4.1771	1.0
3.6003	13.33	40	3.8565	1.0
3.4788	16.67	50	3.5921	1.0
3.4111	20.0	60	3.5524	1.0

Framework versions

Transformers 4.36.0.dev0
Pytorch 2.1.0+cu118
Datasets 2.15.0
Tokenizers 0.15.0