sl-asr-model-primary-small

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

• Loss: 0.3394
• Wer: 0.3198

Model description

Slovenian-ASR-model-primary, 242M paramerters

Intended uses & limitations

Automatic speech recognition for Slovenian

Training and evaluation data

EWCH/data-preprocessed-commonvoice-sl-80s

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.001
• num_epochs: 1
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.3415	1.0	650	0.3394	0.3198

Framework versions

• Transformers 4.39.0.dev0
• Pytorch 2.2.1
• Datasets 2.17.1
• Tokenizers 0.15.2