learning_rate: 0.0003
train_batch_size: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 32
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: cosine
lr_scheduler_warmup_ratio: 0.1
num_epochs: 10
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss
9.1685	0.1953	100	2.1514
2.8539	0.3906	200	0.7851
2.6755	0.5859	300	0.6895
2.7392	0.7812	400	0.7504
2.6266	0.9766	500	0.6679
2.2071	1.1719	600	0.6242
3.162	1.3672	700	0.6029
2.5194	1.5625	800	0.6184
2.4411	1.7578	900	0.5708
2.5937	1.9531	1000	0.5741
2.0553	2.1484	1100	0.5679
2.4125	2.3438	1200	0.5827
1.6075	2.5391	1300	0.5684

Veggissss/ltg_norbert3-base_mask_values

ltg_norbert3-base_mask_values