|
--- |
|
language: |
|
- 'no' |
|
license: apache-2.0 |
|
base_model: NbAiLabBeta/nb-whisper-small |
|
tags: |
|
- audio |
|
- asr |
|
- automatic-speech-recognition |
|
- hf-asr-leaderboard |
|
model-index: |
|
- name: nb-whisper-small-v0.7-semantic |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information Keras had access to. You should |
|
probably proofread and complete it, then remove this comment. --> |
|
|
|
# nb-whisper-small-v0.7-semantic |
|
|
|
This model is a fine-tuned version of [NbAiLabBeta/nb-whisper-small](https://huggingface.co/NbAiLabBeta/nb-whisper-small) on the NbAiLab/ncc_speech_styling_v4 dataset. |
|
It achieves the following results on the evaluation set: |
|
- step: 249 |
|
- validation_nst_loss: 0.6651 |
|
- train_loss: 0.7577 |
|
- validation_nst_wer: 3.1520 |
|
- validation_nst_cer: 0.9723 |
|
- validation_nst_exact_wer: 3.9305 |
|
- validation_nst_exact_cer: 1.0907 |
|
- validation_clean_stortinget_no_loss: 0.6931 |
|
- validation_clean_stortinget_no_wer: 10.0616 |
|
- validation_clean_stortinget_no_cer: 6.2843 |
|
- validation_clean_stortinget_no_exact_wer: 13.4693 |
|
- validation_clean_stortinget_no_exact_cer: 6.8299 |
|
|
|
## 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: 5e-05 |
|
- lr_scheduler_type: linear |
|
- per_device_train_batch_size: 32 |
|
- total_train_batch_size_per_node: 128 |
|
- total_train_batch_size: 1024 |
|
- total_optimization_steps: 250 |
|
- starting_optimization_step: None |
|
- finishing_optimization_step: 250 |
|
- num_train_dataset_workers: 32 |
|
- num_hosts: 8 |
|
- 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 | 0.4428 | 1.3025 | 3.0160 | 0.9359 | 3.7835 | 1.0513 | 0.7127 | 9.9100 | 6.1361 | 13.3223 | 6.6798 | |
|
| 40 | 0.6682 | 0.8769 | 3.4732 | 1.2370 | 4.4423 | 1.3773 | 0.6724 | 10.9593 | 7.0159 | 14.4443 | 7.5606 | |
|
| 80 | 0.6729 | 0.7636 | 3.2664 | 1.0338 | 4.1102 | 1.1575 | 0.6495 | 10.2037 | 6.3692 | 13.5785 | 6.8974 | |
|
| 120 | 0.6729 | 0.7334 | 3.2500 | 1.0310 | 4.0666 | 1.1530 | 0.6760 | 10.1232 | 6.3145 | 13.4646 | 6.8420 | |
|
| 160 | 0.6740 | 0.7517 | 3.1412 | 0.9788 | 3.9033 | 1.0925 | 0.6922 | 10.0734 | 6.2954 | 13.4243 | 6.8245 | |
|
| 200 | 0.6677 | 0.7282 | 3.1575 | 0.9751 | 3.9196 | 1.0898 | 0.6981 | 10.0924 | 6.3054 | 13.4741 | 6.8416 | |
|
| 240 | 0.6658 | 0.7448 | 3.1520 | 0.9919 | 3.9196 | 1.1081 | 0.6929 | 10.0379 | 6.2800 | 13.4599 | 6.8288 | |
|
| 249 | 0.6651 | 0.7577 | 3.1520 | 0.9723 | 3.9305 | 1.0907 | |
|
| 249 | 0.6931 | 0.7577 | 10.0616 | 6.2843 | 13.4693 | 6.8299 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.34.1 |
|
- Datasets 2.15.0 |
|
- Tokenizers 0.14.1 |
|
|
