metadata
language:
- 'no'
license: apache-2.0
tags:
- audio
- asr
- automatic-speech-recognition
- hf-asr-leaderboard
model-index:
- name: scream_sextusdecimus_virtual_tsfix_medium_1e5
results: []
scream_sextusdecimus_virtual_tsfix_medium_1e5
This model is a fine-tuned version of openai/whisper-medium on the NbAiLab/ncc_speech 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: 1e-05
- lr_scheduler_type: linear
- per_device_train_batch_size: 16
- total_train_batch_size_per_node: 64
- total_train_batch_size: 512
- total_optimization_steps: 20,000
- starting_optimization_step: None
- finishing_optimization_step: 20,000
- num_train_dataset_workers: 32
- num_hosts: 8
- total_num_training_examples: 10,240,000
- steps_per_epoch: To be computed after first epoch
- num_beams: None
- dropout: True
- bpe_dropout_probability: 0.1
- activation_dropout_probability: 0.1
Training results
| step | eval_loss | train_loss | eval_wer | eval_cer | eval_exact_wer | eval_exact_cer |
|---|---|---|---|---|---|---|
| 0 | 5.5890 | 2.8362 | 17.4598 | 5.3906 | 17.4598 | 5.3906 |
| 1000 | 5.2798 | 1.0896 | 12.4926 | 3.8321 | 12.4926 | 3.8321 |
| 2000 | 5.2432 | 0.9018 | 11.0351 | 3.9899 | 11.0351 | 3.9899 |
| 3000 | 4.1719 | 0.8159 | 9.8453 | 3.8173 | 9.8453 | 3.8173 |
Framework versions
- Transformers 4.30.0.dev0
- Datasets 2.12.1.dev0
- Tokenizers 0.13.3