learning_rate: 5e-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
21.2588	1.0	97	0.2748	0.1900	0.2906	0.0151	nan
3.1528	2.0	194	0.4417	0.3692	0.4665	-0.9139	0.0542
3.0803	3.0	291	0.3158	0.2011	0.3577	-0.0428	0.3199
16.5636	4.0	388	4.8226	0.2658	0.3713	-0.3780	0.0975
57.3503	5.0	485	5.9922	0.2901	0.4026	-0.5039	0.2237
55.3448	6.0	582	6.0585	0.9760	0.9028	-4.0598	0.2262
48.9687	7.0	679	4.7147	0.1953	0.3231	-0.0123	0.2051
42.0066	8.0	776	4.1447	0.1873	0.3212	0.0290	0.3165
38.9071	9.0	873	3.8776	0.2067	0.3500	-0.0714	0.2643
40.2545	10.0	970	3.7597	0.1909	0.3205	0.0101	0.2581

athirorg/USS-reward-model-exp6-grl-only-without-scaling

USS-reward-model-exp6-grl-only-without-scaling