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 Loss	Epoch	Step	Validation Loss	Model Preparation Time	Wer Ortho	Wer
0.1635	0.1092	32	0.4348	0.0071	32.9116	8.9834
0.1399	0.2184	64	0.4449	0.0071	32.9872	9.6728
0.1381	0.3276	96	0.4426	0.0071	33.1660	10.0987
0.1415	0.4369	128	0.4389	0.0071	33.0766	9.8621
0.1296	0.5461	160	0.3991	0.0071	33.2348	9.6796
0.1353	0.6553	192	0.4331	0.0071	33.0835	9.7066
0.1425	0.7645	224	0.3954	0.0071	32.7465	9.7066
0.1231	0.8737	256	0.4204	0.0071	33.3242	9.5917
0.1316	0.9829	288	0.4196	0.0071	33.1660	9.8824
0.0936	1.0922	320	0.4560	0.0071	33.2210	9.6864
0.0915	1.2014	352	0.4491	0.0071	32.9941	9.5241
0.1247	1.3106	384	0.4414	0.0071	32.8909	9.5512
0.0894	1.4198	416	0.4146	0.0071	32.8428	9.5174
0.0988	1.5290	448	0.4235	0.0071	33.0972	9.5782
0.0926	1.6382	480	0.4400	0.0071	32.9735	9.6999
0.0887	1.7474	512	0.4296	0.0071	32.9459	9.7404
0.089	1.8567	544	0.4248	0.0071	32.8634	9.8080
0.0863	1.9659	576	0.4391	0.0071	32.9391	9.7540
0.0908	2.0751	608	0.4650	0.0071	33.0353	9.6661
0.0746	2.1843	640	0.4523	0.0071	32.8359	9.4768
0.059	2.2935	672	0.4487	0.0071	32.7190	9.1591
0.0796	2.4027	704	0.4498	0.0071	32.5815	8.9361
0.0667	2.5119	736	0.4489	0.0071	32.4577	8.9631
0.0725	2.6212	768	0.4408	0.0071	32.6640	9.1388
0.0577	2.7304	800	0.4448	0.0071	32.5127	9.0172
0.0717	2.8396	832	0.4447	0.0071	32.5334	9.0577
0.0635	2.9488	864	0.4397	0.0071	32.6296	9.0915

sam8000/whisper-large-v3-turbo-finnish-finland-5percent

whisper-large-v3-turbo-finnish-finland-5percent