whisper-small-ru-v7la

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

• Loss: 0.1039
• Wer: 9.5294

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: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 400
• training_steps: 2000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.203	0.1934	200	0.2464	18.6343
0.1693	0.3868	400	0.1949	16.0868
0.1707	0.5803	600	0.1646	14.3649
0.1378	0.7737	800	0.1427	12.8907
0.1194	0.9671	1000	0.1303	11.5698
0.087	1.1605	1200	0.1195	11.1216
0.0824	1.3540	1400	0.1140	10.3550
0.0795	1.5474	1600	0.1083	9.7417
0.0787	1.7408	1800	0.1056	9.5648
0.0781	1.9342	2000	0.1039	9.5294

Framework versions

• Transformers 4.49.0
• Pytorch 2.1.0+cu118
• Datasets 3.3.1
• Tokenizers 0.21.0