This model is a fine-tuned version of openai/whisper-small on the mozilla-foundation/common_voice_13_0 dataset. It achieves the following results on the evaluation set:

Loss: 0.2860
Wer: 19.5690

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: 16
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 500
training_steps: 6000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.5133	0.64	500	0.3286	24.1571
0.3701	1.28	1000	0.2979	22.7444
0.3324	1.92	1500	0.2818	21.2813
0.2016	2.55	2000	0.2802	21.4191
0.1428	3.19	2500	0.2805	20.5921
0.1455	3.83	3000	0.2740	20.3059
0.1162	4.47	3500	0.2797	20.2343
0.0854	5.11	4000	0.2793	19.8526
0.0859	5.75	4500	0.2809	20.0302
0.0574	6.39	5000	0.2852	19.8367
0.0644	7.02	5500	0.2851	19.6644
0.0637	7.66	6000	0.2860	19.5690

Framework versions

Transformers 4.36.0.dev0
Pytorch 2.1.1+cu118
Datasets 2.15.0
Tokenizers 0.15.0

Sleepyp00/whisper-small-Swedish

whisper-small-test2

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

Sleepyp00/
whisper-small-Swedish