metadata

license: apache-2.0
base_model: facebook/wav2vec2-large-xlsr-53
tags:
  - generated_from_trainer
datasets:
  - common_voice
metrics:
  - wer
model-index:
  - name: Model_G_2
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: common_voice
          type: common_voice
          config: id
          split: test
          args: id
        metrics:
          - name: Wer
            type: wer
            value: 0.9848965131456274

Model_G_2

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

Loss: 0.0374
Wer: 0.9849
Cer: 0.7098

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: 500
num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
3.6149	1.07	400	0.4672	1.0114	0.7588
0.4341	2.15	800	0.2008	0.9972	0.7369
0.2665	3.22	1200	0.1283	0.9986	0.7180
0.2114	4.3	1600	0.1016	0.9995	0.7135
0.1768	5.37	2000	0.0774	0.9950	0.7208
0.1531	6.44	2400	0.0682	0.9933	0.7137
0.1352	7.52	2800	0.0690	0.9883	0.7022
0.1252	8.59	3200	0.0656	0.9925	0.7091
0.1144	9.66	3600	0.0521	0.9888	0.7124
0.0986	10.74	4000	0.0527	0.9915	0.7067
0.0875	11.81	4400	0.0531	0.9902	0.7057
0.0883	12.89	4800	0.0488	0.9888	0.7136
0.0812	13.96	5200	0.0461	0.9884	0.7122
0.0721	15.03	5600	0.0474	0.9884	0.7128
0.0681	16.11	6000	0.0469	0.9869	0.7243
0.0671	17.18	6400	0.0450	0.9878	0.7086
0.0613	18.26	6800	0.0492	0.9852	0.7171
0.0573	19.33	7200	0.0435	0.9852	0.7209
0.0531	20.4	7600	0.0389	0.9908	0.7071
0.0493	21.48	8000	0.0423	0.9871	0.7166
0.0477	22.55	8400	0.0416	0.9843	0.7127
0.0441	23.62	8800	0.0372	0.9864	0.7075
0.0412	24.7	9200	0.0408	0.9857	0.7118
0.0392	25.77	9600	0.0407	0.9851	0.7152
0.0359	26.85	10000	0.0383	0.9861	0.7086
0.0347	27.92	10400	0.0373	0.9852	0.7066
0.0327	28.99	10800	0.0374	0.9849	0.7098

Framework versions

Transformers 4.31.0
Pytorch 2.0.1+cu117
Datasets 1.18.3
Tokenizers 0.13.3