learning_rate: 2e-05
train_batch_size: 16
eval_batch_size: 32
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: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 20
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	F1 Macro	Precision	Recall	Accuracy
No log	1.0	354	1.9174	0.3780	0.4605	0.3785	0.5650
2.289	2.0	708	1.2911	0.5305	0.5284	0.5883	0.6926
1.5337	3.0	1062	1.3022	0.5340	0.5466	0.5812	0.7001
1.5337	4.0	1416	1.3496	0.5526	0.5623	0.5737	0.7288
1.0847	5.0	1770	1.4333	0.5665	0.5774	0.5961	0.7439
0.6609	6.0	2124	1.5549	0.5681	0.5782	0.5880	0.7444
0.6609	7.0	2478	1.9182	0.5561	0.5543	0.5754	0.7269
0.388	8.0	2832	2.0856	0.5709	0.5799	0.5793	0.7387
0.1976	9.0	3186	2.4609	0.5564	0.5545	0.5706	0.7368
0.1383	10.0	3540	2.7571	0.5801	0.5833	0.5914	0.7467
0.1383	11.0	3894	3.0987	0.5700	0.5757	0.5711	0.75
0.0618	12.0	4248	3.3352	0.5698	0.5840	0.5633	0.7476

rendchevi/roberta-large-pr_tqacd

roberta-large-pr_tqacd