metadata
language:
- 'no'
license: apache-2.0
base_model: NbAiLab/nb-whisper-large-v3-RC4
tags:
- audio
- asr
- automatic-speech-recognition
- hf-asr-leaderboard
model-index:
- name: nb-whisper-large-v0.8-vad3
results: []
nb-whisper-large-v0.8-vad3
This model is a fine-tuned version of NbAiLab/nb-whisper-large-v3-RC4 on the NbAiLab/ncc_speech_styling_v2_vad3 dataset.
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: 50,000
- starting_optimization_step: None
- finishing_optimization_step: 50,000
- num_train_dataset_workers: 32
- num_hosts: 32
- total_num_training_examples: 51,200,000
- 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 | 0.4259 | 0.9588 | 2.1721 | 0.6246 | 2.7111 | 0.7079 | 0.6807 | 8.5931 | 5.4608 | 11.4221 | 5.8946 |
| 5000 | 0.4376 | 0.5822 | 2.5859 | 0.7793 | 3.0867 | 0.8563 | 0.6738 | 9.1686 | 5.8478 | 12.0792 | 6.3020 |
Framework versions
- Transformers 4.36.2
- Datasets 2.16.1
- Tokenizers 0.15.0