marinone94's picture
update model card README.md
80424d2
|
raw
history blame
3.56 kB
metadata
license: cc0-1.0
tags:
  - automatic-speech-recognition
  - marinone94/nst_sv
  - generated_from_trainer
datasets:
  - nst_sv
model-index:
  - name: ''
    results: []

This model is a fine-tuned version of KBLab/wav2vec2-large-voxrex on the MARINONE94/NST_SV - DISTANT_CHANNEL dataset. It achieves the following results on the evaluation set:

  • Loss: inf
  • Wer: 1.0

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: 0.00075
  • train_batch_size: 32
  • eval_batch_size: 32
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 128
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.02
  • num_epochs: 2.0
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
3.4039 0.05 100 inf 1.0
3.4396 0.11 200 inf 1.0
3.483 0.16 300 inf 1.0
3.5014 0.21 400 inf 1.0
3.331 0.27 500 inf 1.0
3.4809 0.32 600 inf 1.0
3.4678 0.37 700 inf 1.0
3.4596 0.43 800 inf 1.0
3.4644 0.48 900 inf 1.0
3.4671 0.53 1000 inf 1.0
3.6005 0.59 1100 inf 1.0
3.9182 0.64 1200 inf 1.0
3.6466 0.69 1300 inf 1.0
3.6932 0.75 1400 inf 1.0
3.7939 0.8 1500 inf 1.0
3.9284 0.85 1600 inf 1.0
3.7859 0.91 1700 inf 1.0
3.9363 0.96 1800 inf 1.0
3.7573 1.01 1900 inf 1.0
3.7553 1.07 2000 inf 1.0
3.7606 1.12 2100 inf 1.0
3.7514 1.17 2200 inf 1.0
3.7472 1.23 2300 inf 1.0
3.7478 1.28 2400 inf 1.0
3.7496 1.33 2500 inf 1.0
3.7513 1.39 2600 inf 1.0
3.7497 1.44 2700 inf 1.0
3.7539 1.49 2800 inf 1.0
3.7581 1.55 2900 inf 1.0
3.7572 1.6 3000 inf 1.0
3.7589 1.66 3100 inf 1.0
3.7592 1.71 3200 inf 1.0
3.7531 1.76 3300 inf 1.0
3.7567 1.82 3400 inf 1.0
3.7613 1.87 3500 inf 1.0
3.7516 1.92 3600 inf 1.0
3.7581 1.98 3700 inf 1.0

Framework versions

  • Transformers 4.17.0.dev0
  • Pytorch 1.10.2+cu102
  • Datasets 1.18.3
  • Tokenizers 0.11.0