metadata

language:
  - hu
license: apache-2.0
tags:
  - automatic-speech-recognition
  - generated_from_trainer
  - hf-asr-leaderboard
  - hu
  - model_for_talk
  - mozilla-foundation/common_voice_7_0
  - robust-speech-event
datasets:
  - mozilla-foundation/common_voice_7_0
model-index:
  - name: XLS-R-300M - Hungarian
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Common Voice 7
          type: mozilla-foundation/common_voice_7_0
          args: hu
        metrics:
          - name: Test WER
            type: wer
            value: 31.099
          - name: Test CER
            type: cer
            value: 6.737
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Robust Speech Event - Dev Data
          type: speech-recognition-community-v2/dev_data
          args: hu
        metrics:
          - name: Test WER
            type: wer
            value: 45.469
          - name: Test CER
            type: cer
            value: 15.727
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Robust Speech Event - Test Data
          type: speech-recognition-community-v2/eval_data
          args: hu
        metrics:
          - name: Test WER
            type: wer
            value: 48.2

wav2vec2-large-xls-r-300m-hungarian

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the MOZILLA-FOUNDATION/COMMON_VOICE_7_0 - HU dataset. It achieves the following results on the evaluation set:

Loss: 0.2562
Wer: 0.3112

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: 7e-05
train_batch_size: 32
eval_batch_size: 32
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 1000
num_epochs: 50.0
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
2.3964	3.52	1000	1.2251	0.8781
1.3176	7.04	2000	0.3872	0.4462
1.1999	10.56	3000	0.3244	0.3922
1.1633	14.08	4000	0.3014	0.3704
1.1132	17.61	5000	0.2913	0.3623
1.0888	21.13	6000	0.2864	0.3498
1.0487	24.65	7000	0.2821	0.3435
1.0431	28.17	8000	0.2739	0.3308
0.9896	31.69	9000	0.2629	0.3243
0.9839	35.21	10000	0.2806	0.3308
0.9586	38.73	11000	0.2650	0.3235
0.9501	42.25	12000	0.2585	0.3173
0.938	45.77	13000	0.2561	0.3117
0.921	49.3	14000	0.2559	0.3115

Framework versions

Transformers 4.16.0.dev0
Pytorch 1.10.1+cu102
Datasets 1.17.1.dev0
Tokenizers 0.11.0