learning_rate: 1e-05
train_batch_size: 16
eval_batch_size: 8
seed: 42
optimizer: Use OptimizerNames.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: 500
training_steps: 5000
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer
0.4317	0.2759	500	0.6019	68.1049
0.5176	0.5519	1000	0.5323	62.2702
0.5129	0.8278	1500	0.4980	60.6521
0.2863	1.1038	2000	0.4851	57.6857
0.3048	1.3797	2500	0.4792	58.0780
0.3355	1.6556	3000	0.4662	55.5038
0.3752	1.9316	3500	0.4611	55.2096
0.1864	2.2075	4000	0.4689	55.4793
0.1938	2.4834	4500	0.4691	54.8909
0.2211	2.7594	5000	0.4667	54.1799

bumpingbell/whisper-small-trained_20250226

whisper-small-trained_20250226