learning_rate: 5e-05
train_batch_size: 500
eval_batch_size: 500
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
num_epochs: 10
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1 score
0.0224	1.0	2376	0.0209	0.9915	0.9310	0.9308	0.9298
0.0203	2.0	4752	0.0198	0.9918	0.9308	0.9306	0.9289
0.0163	3.0	7128	0.0186	0.9919	0.9612	0.9423	0.9468
0.0168	4.0	9504	0.0195	0.9919	0.9351	0.9200	0.9245
0.0132	5.0	11880	0.0170	0.9923	0.9625	0.9595	0.9603
0.0154	6.0	14256	0.0169	0.9924	0.9694	0.9603	0.9630
0.0105	7.0	16632	0.0165	0.9925	0.9701	0.9590	0.9627
0.0133	8.0	19008	0.0161	0.9927	0.9661	0.9663	0.9650
0.013	9.0	21384	0.0158	0.9927	0.9782	0.9677	0.9711
0.0112	10.0	23760	0.0156	0.9927	0.9631	0.9678	0.9646

gates04/Structured-FP16-KD-SD-NID

Structured-FP16-KD-SD-NID