birgermoell/whisper-tiny-sv-fast

Whisper Tiny Swedish Fast

This model is a fine-tuned version of openai/whisper-tiny on the mozilla-foundation/common_voice_11_0 sv-SE dataset. It achieves the following results on the evaluation set:

• Loss: 1.4344
• Wer: 73.0163

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
• train_batch_size: 128
• eval_batch_size: 64
• 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
1.5547	6.01	1000	1.9244	113.4448
0.7244	12.02	2000	1.4593	81.0128
0.3583	18.03	3000	1.4019	74.3415
0.2157	25.01	4000	1.4249	73.8953
0.1897	31.02	5000	1.4344	73.0163

Framework versions

• Transformers 4.26.0.dev0
• Pytorch 1.13.0+cu117
• Datasets 2.7.1.dev0
• Tokenizers 0.13.2