learning_rate: 1e-05
train_batch_size: 16
eval_batch_size: 16
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 32
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: 0.1
num_epochs: 10
mixed_precision_training: Native AMP
label_smoothing_factor: 0.1

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
1.7036	1.0	822	0.9692	0.5957	0.5947
1.4881	2.0	1644	0.8902	0.6363	0.6327
1.1695	3.0	2466	0.9503	0.6452	0.6450
1.0144	4.0	3288	1.0459	0.6495	0.6491
0.9116	5.0	4110	1.0702	0.6433	0.6427
0.8019	6.0	4932	1.1632	0.6535	0.6534
0.7347	7.0	5754	1.2237	0.6504	0.6504
0.7064	8.0	6576	1.2701	0.6516	0.6517
0.6902	9.0	7398	1.2603	0.6488	0.6489

Inori612/29_NEW_phobert-large

29_NEW_phobert-large