metadata

base_model: facebook/mms-1b-all
datasets:
  - common_voice_17_0
library_name: transformers
license: cc-by-nc-4.0
metrics:
  - wer
  - bleu
tags:
  - generated_from_trainer
model-index:
  - name: wav2vec2-mms-1b-CV17.0-training_set_variations
    results:
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: common_voice_17_0
          type: common_voice_17_0
          config: ta
          split: validation
          args: ta
        metrics:
          - type: wer
            value: 0.3699525493114998
            name: Wer
          - type: bleu
            value: 0.4072321954028345
            name: Bleu

wav2vec2-mms-1b-CV17.0-training_set_variations

This model is a fine-tuned version of facebook/mms-1b-all on the common_voice_17_0 dataset. It achieves the following results on the evaluation set:

Loss: 0.2281
Wer: 0.3700
Cer: 0.0598
Bleu: 0.4072

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: 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_ratio: 0.15
training_steps: 2000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer	Bleu
7.0089	1.5625	100	0.2991	0.4260	0.0693	0.3354
0.2087	3.125	200	0.2305	0.3968	0.0634	0.3678
0.1924	4.6875	300	0.2291	0.3879	0.0624	0.3799
0.1799	6.25	400	0.2290	0.3859	0.0629	0.3830
0.1698	7.8125	500	0.2224	0.3700	0.0600	0.4119
0.1587	9.375	600	0.2246	0.3672	0.0601	0.4129
0.1547	10.9375	700	0.2176	0.3855	0.0604	0.3820
0.1446	12.5	800	0.2273	0.3907	0.0619	0.3755
0.1404	14.0625	900	0.2239	0.3713	0.0605	0.4035
0.1333	15.625	1000	0.2261	0.3699	0.0602	0.4123
0.1251	17.1875	1100	0.2281	0.3700	0.0598	0.4072

Framework versions

Transformers 4.44.2
Pytorch 2.4.1+cu121
Datasets 3.0.0
Tokenizers 0.19.1