learning_rate: 1e-05
train_batch_size: 8
eval_batch_size: 4
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 32
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: 300
training_steps: 3000
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer
3.076	0.0833	250	1.2445	63.2079
0.9739	0.1667	500	0.8483	49.4433
0.8257	0.25	750	0.7639	43.0172
0.7021	0.3333	1000	0.7260	38.7474
0.6603	1.0833	1250	0.6980	37.8832
0.6147	1.1667	1500	0.6994	39.7526
0.5725	1.25	1750	0.6786	38.3574
0.5129	1.3333	2000	0.6728	37.7148

wwwtwwwt/whisper-tiny-ScienceTechnology

whisper-tiny-ScienceTechnology