Hugging Face's logo

ngohuudang
/
test_bug2

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
test_bug2 / README.md
ngohuudang's picture
ngohuudang
update model card README.md
10a8abc
preview code
|
raw
history blame
3.63 kB
metadata
license: cc-by-nc-4.0
tags:
  - generated_from_trainer
model-index:
  - name: test_bug2
    results: []

test_bug2

This model is a fine-tuned version of nguyenvulebinh/wav2vec2-base-vietnamese-250h on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.2977
  • Wer: 0.1839

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: 4
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 16
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • num_epochs: 10
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
1.1568 0.27 50 0.2764 0.1985
0.0979 0.53 100 0.2421 0.1813
0.1018 0.8 150 0.2420 0.1809
0.1011 1.07 200 0.2520 0.1992
0.0947 1.34 250 0.2580 0.1885
0.1077 1.6 300 0.2641 0.2001
0.109 1.87 350 0.3196 0.2156
0.1239 2.14 400 0.3298 0.2163
0.1286 2.41 450 0.3392 0.2436
0.1515 2.67 500 0.3821 0.2450
0.157 2.94 550 0.3771 0.2521
0.1296 3.21 600 0.3917 0.2541
0.1351 3.48 650 0.3670 0.2366
0.1387 3.74 700 0.3503 0.2347
0.1336 4.01 750 0.4018 0.2627
0.114 4.28 800 0.3699 0.2723
0.1254 4.54 850 0.3395 0.2404
0.119 4.81 900 0.3410 0.2340
0.1 5.08 950 0.3302 0.2216
0.0968 5.35 1000 0.3346 0.2255
0.0965 5.61 1050 0.3144 0.2140
0.0906 5.88 1100 0.3277 0.2109
0.0968 6.15 1150 0.3300 0.2141
0.0818 6.42 1200 0.3272 0.2085
0.0836 6.68 1250 0.3177 0.2014
0.0803 6.95 1300 0.3185 0.2005
0.0727 7.22 1350 0.3110 0.1928
0.0687 7.49 1400 0.3118 0.1965
0.0698 7.75 1450 0.3170 0.1955
0.0651 8.02 1500 0.3119 0.1929
0.0648 8.29 1550 0.3058 0.1904
0.0612 8.56 1600 0.3087 0.1935
0.0578 8.82 1650 0.3076 0.1871
0.0557 9.09 1700 0.3037 0.1862
0.0542 9.36 1750 0.2990 0.1858
0.0551 9.62 1800 0.2962 0.1837
0.0514 9.89 1850 0.2977 0.1839

Framework versions

  • Transformers 4.16.0
  • Pytorch 1.13.1+cu116
  • Datasets 1.18.3
  • Tokenizers 0.12.1