whisper-small-ISSAI_KSC_335RS

This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1140
• Wer: 14.5474
• Cer: 3.3911

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: linear
• lr_scheduler_warmup_steps: 100
• training_steps: 5001
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.0918	0.11	500	0.1361	17.5287	4.1649
0.0795	0.22	1000	0.1327	16.8463	4.0455
0.0783	0.33	1500	0.1311	16.7385	3.9758
0.0699	0.43	2000	0.1273	16.5050	3.9360
0.0705	0.54	2500	0.1230	15.8944	3.8489
0.0622	0.65	3000	0.1194	16.2177	3.8514
0.0642	0.76	3500	0.1157	14.7989	3.5230
0.0697	0.87	4000	0.1152	15.1221	3.5802
0.0658	0.98	4500	0.1121	14.6731	3.4608
0.0271	1.09	5000	0.1140	14.5474	3.3911

Framework versions

• Transformers 4.27.0.dev0
• Pytorch 1.13.1+cu116
• Datasets 2.10.1
• Tokenizers 0.13.2