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: 0.45588302699729566
            name: Wer
          - type: bleu
            value: 0.30120570288375
            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: 0.4242
Wer: 0.4559
Cer: 0.0764
Bleu: 0.3012

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
13.1596	12.5	50	6.8694	1.0	0.9625	0.0
3.2131	25.0	100	0.4085	0.4707	0.0784	0.2830
0.1719	37.5	150	0.2583	0.3920	0.0650	0.3818
0.0962	50.0	200	0.2869	0.4118	0.0682	0.3547
0.0648	62.5	250	0.3209	0.4213	0.0696	0.3435
0.0613	75.0	300	0.3404	0.4454	0.0742	0.3200
0.0515	87.5	350	0.3744	0.4385	0.0734	0.3289
0.0426	100.0	400	0.3835	0.4479	0.0748	0.3078
0.0384	112.5	450	0.3776	0.4432	0.0746	0.3243
0.0363	125.0	500	0.4053	0.4371	0.0732	0.3251
0.0385	137.5	550	0.4225	0.4520	0.0772	0.3115
0.0343	150.0	600	0.4295	0.4463	0.0758	0.3167
0.0371	162.5	650	0.4242	0.4559	0.0764	0.3012

Framework versions

Transformers 4.44.2
Pytorch 2.4.1+cu121
Datasets 3.0.0
Tokenizers 0.19.1