learning_rate: 2e-05
train_batch_size: 2
eval_batch_size: 1
seed: 42
distributed_type: multi-GPU
gradient_accumulation_steps: 10
total_train_batch_size: 20
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
num_epochs: 10
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Mse	Mae	R2	Spearman Correlation
5.9160	1.0	97	0.5358	0.2765	0.3934	-0.4332	-0.0216
5.9268	2.0	194	0.5358	0.2765	0.3934	-0.4332	-0.0216
5.9105	3.0	291	0.5358	0.2765	0.3934	-0.4332	-0.0216
5.9198	4.0	388	0.5358	0.2765	0.3934	-0.4332	-0.0216
5.9129	5.0	485	0.5358	0.2765	0.3934	-0.4332	-0.0216
5.9200	6.0	582	0.5358	0.2765	0.3934	-0.4332	-0.0216
5.9228	7.0	679	0.5358	0.2765	0.3934	-0.4332	-0.0216
5.9265	8.0	776	0.5358	0.2765	0.3934	-0.4332	-0.0216
5.9074	9.0	873	0.5358	0.2765	0.3934	-0.4332	-0.0216
5.9090	10.0	970	0.5358	0.2765	0.3934	-0.4332	-0.0216

athirorg/USS-reward-model-hybrid-with-history-alpha1.0

USS-reward-model-hybrid-with-history-alpha1.0