versae/whisper-large-nob-ncc-s-lr5e-6

Whisper Large Norwegian

This model is a fine-tuned version of openai/whisper-large-v2 on the NbAiLab/NCC_S dataset. It achieves the following results on the evaluation set:

• Loss: 0.2784
• Wer: 12.0585

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-06
• train_batch_size: 12
• eval_batch_size: 6
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.6755	0.2	1000	0.3108	14.3118
0.673	0.4	2000	0.3004	13.4592
0.6378	0.6	3000	0.2865	13.0024
0.5776	0.8	4000	0.2809	12.6675
0.5962	1.0	5000	0.2784	12.0585

Framework versions

• Transformers 4.26.0.dev0
• Pytorch 1.13.0+cu117
• Datasets 2.7.1.dev0
• Tokenizers 0.11.0