metadata
language:
- 'no'
license: apache-2.0
base_model: NbAiLab/nb-whisper-large-v0.7
tags:
- audio
- asr
- automatic-speech-recognition
- hf-asr-leaderboard
model-index:
- name: nb-whisper-large-v0.7-semantic
results: []
nb-whisper-large-v0.7-semantic
This model is a fine-tuned version of NbAiLab/nb-whisper-large-v0.7 on the NbAiLab/ncc_speech_styling_v4 dataset. It achieves the following results on the evaluation set:
- step: 249
- validation_nst_loss: 0.5702
- train_loss: 0.7256
- validation_nst_wer: 2.2157
- validation_nst_cer: 0.7047
- validation_nst_exact_wer: 2.7819
- validation_nst_exact_cer: 0.7949
- validation_clean_stortinget_no_loss: 0.6889
- validation_clean_stortinget_no_wer: 9.2705
- validation_clean_stortinget_no_cer: 6.0401
- validation_clean_stortinget_no_exact_wer: 12.1907
- validation_clean_stortinget_no_exact_cer: 6.5036
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
- lr_scheduler_type: linear
- per_device_train_batch_size: 8
- total_train_batch_size_per_node: 32
- total_train_batch_size: 1024
- total_optimization_steps: 250
- starting_optimization_step: None
- finishing_optimization_step: 250
- num_train_dataset_workers: 32
- num_hosts: 32
- total_num_training_examples: 256,000
- steps_per_epoch: To be computed after first epoch
- num_beams: None
- weight_decay: 0.01
- adam_beta1: 0.9
- adam_beta2: 0.98
- adam_epsilon: 1e-06
- dropout: True
- bpe_dropout_probability: 0.2
- activation_dropout_probability: 0.1
Training results
| step | validation_nst_loss | train_loss | validation_nst_wer | validation_nst_cer | validation_nst_exact_wer | validation_nst_exact_cer | validation_clean_stortinget_no_loss | validation_clean_stortinget_no_wer | validation_clean_stortinget_no_cer | validation_clean_stortinget_no_exact_wer | validation_clean_stortinget_no_exact_cer |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 0.4293 | 1.1323 | 2.1830 | 0.6702 | 2.7220 | 0.7519 | 0.7055 | 8.9815 | 5.7582 | 11.8396 | 6.2120 |
| 40 | 0.6010 | 0.7610 | 2.2973 | 0.7215 | 2.9996 | 0.8297 | 0.6828 | 9.5026 | 6.1967 | 12.4635 | 6.6537 |
| 80 | 0.5735 | 0.7486 | 2.2973 | 0.7327 | 2.8690 | 0.8224 | 0.6801 | 9.3510 | 6.0762 | 12.2809 | 6.5375 |
| 120 | 0.5715 | 0.7146 | 2.2266 | 0.7038 | 2.8200 | 0.7967 | 0.6846 | 9.4008 | 6.1261 | 12.2975 | 6.5804 |
| 160 | 0.5733 | 0.7211 | 2.2211 | 0.7094 | 2.8091 | 0.8031 | 0.6925 | 9.3108 | 6.0520 | 12.2239 | 6.5134 |
| 200 | 0.5703 | 0.7194 | 2.2157 | 0.7057 | 2.7601 | 0.7921 | 0.6905 | 9.2326 | 6.0104 | 12.1623 | 6.4724 |
| 240 | 0.5691 | 0.7292 | 2.2157 | 0.7057 | 2.7710 | 0.7931 | 0.6885 | 9.2729 | 6.0401 | 12.1860 | 6.5013 |
| 250 | 0.5702 | 0.7256 | 2.2157 | 0.7047 | 2.7819 | 0.7949 | 0.6889 | 9.2705 | 6.0401 | 12.1907 | 6.5036 |
Framework versions
- Transformers 4.35.2
- Datasets 2.15.0
- Tokenizers 0.14.1