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: 0.41655589677774213
            name: Wer
          - type: bleu
            value: 0.3602695476100989
            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: 0.3166
Wer: 0.4166
Cer: 0.0689
Bleu: 0.3603

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
11.8671	3.125	50	3.8104	1.0011	0.9249	0.0
1.5965	6.25	100	0.2796	0.4122	0.0680	0.3517
0.2054	9.375	150	0.2320	0.3751	0.0620	0.4040
0.1702	12.5	200	0.2367	0.3794	0.0633	0.3939
0.1495	15.625	250	0.2527	0.4168	0.0680	0.3457
0.1457	18.75	300	0.2536	0.3973	0.0662	0.3723
0.1264	21.875	350	0.2765	0.4175	0.0693	0.3546
0.1092	25.0	400	0.2711	0.4032	0.0673	0.3701
0.0952	28.125	450	0.2828	0.4139	0.0691	0.3605
0.0919	31.25	500	0.2972	0.4283	0.0721	0.3355
0.0909	34.375	550	0.2971	0.4155	0.0686	0.3567
0.0744	37.5	600	0.3086	0.4241	0.0705	0.3461
0.069	40.625	650	0.3166	0.4166	0.0689	0.3603

Framework versions

Transformers 4.44.2
Pytorch 2.4.1+cu121
Datasets 3.0.0
Tokenizers 0.19.1