umutmb/whisper-small-sl

Whisper Small Slovenian

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

• Loss: 0.7109
• Wer: 27.0262

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: 32
• eval_batch_size: 32
• 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

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.006	13.0	1000	0.5072	28.3382
0.0011	26.0	2000	0.7109	27.0262
0.0005	39.01	3000	0.8865	27.1866
0.0	52.01	4000	1.0261	29.0087
0.0	65.01	5000	1.0872	28.9650

Framework versions

• Transformers 4.31.0.dev0
• Pytorch 2.0.1+cu117
• Datasets 2.13.2.dev0
• Tokenizers 0.13.3