README.md

metadata

tags:
  - generated_from_trainer
datasets:
  - evanarlian/common_voice_11_0_id_filtered
metrics:
  - wer
model-index:
  - name: wav2vec2-xls-r-113m-id
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: evanarlian/common_voice_11_0_id_filtered
          type: evanarlian/common_voice_11_0_id_filtered
        metrics:
          - name: Wer
            type: wer
            value: 0.39516649755557604

wav2vec2-xls-r-113m-id

This model is a fine-tuned version of evanarlian/distil-wav2vec2-xls-r-113m-id on the evanarlian/common_voice_11_0_id_filtered dataset. It achieves the following results on the evaluation set:

• Loss: 0.3280
• Wer: 0.3952

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: 0.0002
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 3
• total_train_batch_size: 24
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.3
• num_epochs: 25.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
3.2512	0.92	1000	2.9098	1.0000
2.163	1.84	2000	1.4810	0.9941
1.2472	2.75	3000	0.9604	0.9196
1.0166	3.67	4000	0.8240	0.8498
0.8765	4.59	5000	0.6873	0.7741
0.7712	5.51	6000	0.6083	0.7111
0.6892	6.43	7000	0.5546	0.6592
0.6314	7.35	8000	0.5022	0.6108
0.5779	8.26	9000	0.4850	0.5825
0.5245	9.18	10000	0.4665	0.5538
0.4858	10.1	11000	0.4282	0.5279
0.4616	11.02	12000	0.4053	0.5082
0.421	11.94	13000	0.3809	0.4935
0.4064	12.86	14000	0.3706	0.4781
0.3758	13.77	15000	0.3743	0.4672
0.3598	14.69	16000	0.3571	0.4521
0.3441	15.61	17000	0.3455	0.4368
0.3279	16.53	18000	0.3398	0.4386
0.3086	17.45	19000	0.3512	0.4284
0.3013	18.37	20000	0.3321	0.4233
0.2963	19.28	21000	0.3391	0.4178
0.2831	20.2	22000	0.3438	0.4114
0.2801	21.12	23000	0.3336	0.4056
0.2623	22.04	24000	0.3317	0.4012
0.263	22.96	25000	0.3280	0.4005
0.2529	23.88	26000	0.3268	0.3951
0.2492	24.79	27000	0.3280	0.3952

Framework versions

• Transformers 4.25.1
• Pytorch 1.12.1
• Datasets 2.7.1
• Tokenizers 0.13.1