whisper-large-v3-turbo-icelandic-iceland-0.11-cer-filtered

This model is a fine-tuned version of openai/whisper-large-v3-turbo on the fleurs dataset. It achieves the following results on the evaluation set:

• Loss: 0.4170
• Model Preparation Time: 0.0065
• Wer Ortho: 25.8789
• Wer: 13.6496

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: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 128
• 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_ratio: 0.06
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time	Wer Ortho	Wer
0.2952	0.1017	32	0.5322	0.0065	29.1992	18.2962
0.2454	0.2035	64	0.5165	0.0065	30.2734	19.7483
0.2157	0.3052	96	0.5138	0.0065	29.5898	18.4898
0.2006	0.4070	128	0.5135	0.0065	29.9805	18.2962
0.1753	0.5087	160	0.5199	0.0065	29.4922	18.0058
0.1728	0.6105	192	0.4628	0.0065	29.7852	18.1026
0.1718	0.7122	224	0.4562	0.0065	28.7109	16.6505
0.1631	0.8140	256	0.4364	0.0065	29.5898	18.2962
0.1614	0.9157	288	0.4461	0.0065	28.8086	16.9409
0.1295	1.0159	320	0.4348	0.0065	27.5391	15.3921
0.1131	1.1176	352	0.4490	0.0065	28.125	15.9729
0.1104	1.2194	384	0.4377	0.0065	27.6367	16.0697
0.1192	1.3211	416	0.4263	0.0065	27.2461	15.1985
0.1139	1.4229	448	0.4183	0.0065	27.5391	15.6825
0.1085	1.5246	480	0.4321	0.0065	28.2227	16.1665
0.1122	1.6264	512	0.4482	0.0065	27.0508	15.0048
0.1092	1.7281	544	0.4115	0.0065	26.9531	14.9080
0.1091	1.8299	576	0.4246	0.0065	26.6602	14.8112
0.1065	1.9316	608	0.4372	0.0065	27.7344	16.2633
0.0785	2.0318	640	0.4554	0.0065	26.9531	14.8112
0.0768	2.1335	672	0.4434	0.0065	26.3672	14.3272
0.0749	2.2353	704	0.4343	0.0065	27.8320	15.7793
0.0777	2.3370	736	0.4170	0.0065	25.8789	13.6496
0.0762	2.4388	768	0.4372	0.0065	26.8555	14.3272
0.0756	2.5405	800	0.4403	0.0065	26.4648	14.4240
0.0785	2.6423	832	0.4276	0.0065	27.0508	14.6176
0.0734	2.7440	864	0.4328	0.0065	25.9766	13.6496
0.0772	2.8458	896	0.4301	0.0065	26.2695	13.8432
0.0724	2.9475	928	0.4364	0.0065	25.9766	13.6496

Framework versions

• Transformers 4.51.3
• Pytorch 2.5.1
• Datasets 3.6.0
• Tokenizers 0.21.1