Revolt47's picture
End of training
ad4eea1
metadata
license: apache-2.0
base_model: facebook/wav2vec2-base-960h
tags:
  - generated_from_trainer
metrics:
  - wer
model-index:
  - name: wav2vec2-urdufinetuned
    results: []

wav2vec2-urdufinetuned

This model is a fine-tuned version of facebook/wav2vec2-base-960h on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 3.6089
  • 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: 0.001
  • train_batch_size: 10
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 100
  • num_epochs: 4
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
10.903 0.06 100 3.7302 1.0
3.6693 0.11 200 3.6193 1.0
3.6908 0.17 300 3.6678 1.0
3.6565 0.22 400 3.6365 1.0
3.6348 0.28 500 3.6443 1.0
3.6878 0.33 600 3.6583 1.0
3.572 0.39 700 3.6304 1.0
3.6749 0.44 800 3.6420 1.0
3.6872 0.5 900 3.6469 1.0
3.6594 0.56 1000 3.6278 1.0
3.6131 0.61 1100 3.6169 1.0
3.5748 0.67 1200 3.6234 1.0
3.6181 0.72 1300 3.6494 1.0
3.6164 0.78 1400 3.6248 1.0
3.6688 0.83 1500 3.6610 1.0
4.1978 0.89 1600 3.6903 1.0
3.7485 0.94 1700 3.6275 1.0
3.649 1.0 1800 3.6139 1.0
3.5834 1.06 1900 3.6161 1.0
3.6338 1.11 2000 3.6647 1.0
3.5427 1.17 2100 3.6129 1.0
3.6117 1.22 2200 3.6084 1.0
3.6726 1.28 2300 3.6149 1.0
3.6278 1.33 2400 3.6342 1.0
3.6746 1.39 2500 3.6102 1.0
3.574 1.44 2600 3.7048 1.0
3.5892 1.5 2700 3.6126 1.0
3.6575 1.56 2800 3.6163 1.0
3.592 1.61 2900 3.6610 1.0
3.6506 1.67 3000 3.6127 1.0
3.5823 1.72 3100 3.6071 1.0
3.6674 1.78 3200 3.6032 1.0
3.6017 1.83 3300 3.6236 1.0
3.5865 1.89 3400 3.6208 1.0
3.646 1.94 3500 3.6074 1.0
3.6042 2.0 3600 3.6442 1.0
3.56 2.06 3700 3.6076 1.0
3.6241 2.11 3800 3.6051 1.0
3.6245 2.17 3900 3.6074 1.0
3.5764 2.22 4000 3.6238 1.0
3.6168 2.28 4100 3.6192 1.0
3.6143 2.33 4200 3.6093 1.0
3.613 2.39 4300 3.6123 1.0
3.6178 2.44 4400 3.6135 1.0
3.6234 2.5 4500 3.6161 1.0
3.5833 2.56 4600 3.6064 1.0
3.5759 2.61 4700 3.6077 1.0
3.6747 2.67 4800 3.6123 1.0
3.5914 2.72 4900 3.6041 1.0
3.6342 2.78 5000 3.6208 1.0
3.5883 2.83 5100 3.6056 1.0
3.5563 2.89 5200 3.6159 1.0
3.6213 2.94 5300 3.6173 1.0
3.6507 3.0 5400 3.6031 1.0
3.549 3.06 5500 3.6371 1.0
3.5712 3.11 5600 3.6049 1.0
3.5731 3.17 5700 3.6273 1.0
3.6232 3.22 5800 3.6012 1.0
3.6406 3.28 5900 3.6020 1.0
3.6456 3.33 6000 3.6015 1.0
3.6268 3.39 6100 3.6047 1.0
3.6286 3.44 6200 3.6023 1.0
3.609 3.5 6300 3.6053 1.0
3.6256 3.56 6400 3.6040 1.0
3.5537 3.61 6500 3.6075 1.0
3.5214 3.67 6600 3.6055 1.0
3.6031 3.72 6700 3.6156 1.0
3.6624 3.78 6800 3.6037 1.0
3.5813 3.83 6900 3.6030 1.0
3.6514 3.89 7000 3.6043 1.0
3.5535 3.94 7100 3.6091 1.0
3.5954 4.0 7200 3.6089 1.0

Framework versions

  • Transformers 4.36.0.dev0
  • Pytorch 2.1.0+cu118
  • Datasets 2.14.6
  • Tokenizers 0.14.1