learning_rate: 0.0001
train_batch_size: 8
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: linear
lr_scheduler_warmup_steps: 0.1
num_epochs: 4
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Exact Match	F1
2.4228	0.2818	2500	2.1270	38.8883	52.5500
1.6616	0.5636	5000	1.4774	55.3014	69.3763
1.4681	0.8455	7500	1.3950	57.3259	72.3313
1.3595	1.1273	10000	1.3410	59.7278	73.9674
1.3782	1.4091	12500	1.2437	60.2997	74.3828
1.3025	1.6909	15000	1.2301	61.2833	75.4122
1.2383	1.9727	17500	1.2200	62.1526	76.3566
1.1704	2.2545	20000	1.1993	62.4042	76.6643
1.2248	2.5364	22500	1.1663	62.5186	76.8061
1.2025	2.8182	25000	1.1687	63.0562	77.1522
1.0719	3.1000	27500	1.1916	63.2163	77.4445
1.1436	3.3818	30000	1.1340	63.5251	77.6837
1.1182	3.6636	32500	1.1377	63.7195	78.0391
1.1468	3.9454	35000	1.1263	63.7081	77.9738
1.1468	4.0	35484	1.1275	63.7310	77.9541

NastasiaM/mbert-loraxs-qa-LTr64-qkvd-iscale-0.7-gain1.2-A-norm

mbert-loraxs-qa-LTr64-qkvd-iscale-0.7-gain1.2-A-norm