README.md · vutankiet2901/wav2vec2-xls-r-1b-ja at main

metadata

license: apache-2.0
language:
  - ja
tags:
  - automatic-speech-recognition
  - common-voice
  - hf-asr-leaderboard
  - ja
  - robust-speech-event
datasets:
  - mozilla-foundation/common_voice_8_0
model-index:
  - name: wav2vec2-xls-r-1b
    results:
      - task:
          name: Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Common Voice 7.0
          type: mozilla-foundation/common_voice_7_0
          args: ja
        metrics:
          - name: Test WER (with LM)
            type: wer
            value: 11.77
          - name: Test CER (with LM)
            type: cer
            value: 5.22
      - task:
          name: Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Common Voice 8.0
          type: mozilla-foundation/common_voice_8_0
          args: ja
        metrics:
          - name: Test WER (with LM)
            type: wer
            value: 12.23
          - name: Test CER (with LM)
            type: cer
            value: 5.33
      - task:
          name: Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Robust Speech Event - Dev Data
          type: speech-recognition-community-v2/dev_data
          args: ja
        metrics:
          - name: Test WER (with LM)
            type: wer
            value: 29.35
          - name: Test CER (with LM)
            type: cer
            value: 16.43
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Robust Speech Event - Test Data
          type: speech-recognition-community-v2/eval_data
          args: ja
        metrics:
          - name: Test CER
            type: cer
            value: 19.48

Model description

This model is a fine-tuned version of facebook/wav2vec2-xls-r-1b on the MOZILLA-FOUNDATION/COMMON_VOICE_8_0 - JA

Benchmark WER result:

	COMMON VOICE 7.0	COMMON VOICE 8.0
without LM	16.97	17.95
with 4-grams LM	11.77	12.23

Benchmark CER result:

	COMMON VOICE 7.0	COMMON VOICE 8.0
without LM	6.82	7.05
with 4-grams LM	5.22	5.33

Evaluation

Please use the eval.py file to run the evaluation:

pip install mecab-python3 unidic-lite pykakasi
python eval.py --model_id vutankiet2901/wav2vec2-xls-r-1b-ja --dataset mozilla-foundation/common_voice_8_0 --config ja --split test --log_outputs

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 5e-05
train_batch_size: 16
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 64
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 2000
num_epochs: 100.0
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
3.484	9.49	1500	1.1849	0.7543	0.4099
1.3582	18.98	3000	0.4320	0.3489	0.1591
1.1716	28.48	4500	0.3835	0.3175	0.1454
1.0951	37.97	6000	0.3732	0.3033	0.1405
1.04	47.47	7500	0.3485	0.2898	0.1360
0.9768	56.96	9000	0.3386	0.2787	0.1309
0.9129	66.45	10500	0.3363	0.2711	0.1272
0.8614	75.94	12000	0.3386	0.2676	0.1260
0.8092	85.44	13500	0.3356	0.2610	0.1240
0.7658	94.93	15000	0.3316	0.2564	0.1218

Framework versions

Transformers 4.16.0.dev0
Pytorch 1.10.1+cu102
Datasets 1.18.3
Tokenizers 0.11.0