learning_rate: 5e-05
train_batch_size: 4
eval_batch_size: 8
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 250
num_epochs: 10
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss
2.2978	0.8	200	2.5087
2.3341	1.6	400	2.3892
2.3159	2.4	600	2.3218
2.2470	3.2	800	2.3972
2.2456	4.0	1000	2.2644
2.1870	4.8	1200	2.2401
2.0843	5.6	1400	2.2379
2.1481	6.4	1600	2.2102
2.0924	7.2	1800	2.2164
2.0259	8.0	2000	2.2231
2.0335	8.8	2200	2.1902
2.0944	9.6	2400	2.1871
2.0179	10.0	2500	2.1875

Pealik/detr-fashionpedia

detr-fashionpedia