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
2.2944	1.0	591	0.1483	0.1483	0.2698	-0.3561	nan
1.1189	2.0	1182	0.1079	0.1079	0.1692	0.0130	nan
1.0480	3.0	1773	0.1327	0.1327	0.2664	-0.2131	nan
1.0150	4.0	2364	0.1076	0.1076	0.1544	0.0162	nan
0.9418	5.0	2955	0.1063	0.1063	0.1511	0.0278	-0.0018
0.8678	6.0	3546	0.1120	0.1120	0.2055	-0.0243	-0.0018
0.7709	7.0	4137	0.1102	0.1102	0.1855	-0.0080	-0.0037
0.6529	8.0	4728	0.1091	0.1091	0.1804	0.0022	0.0301
0.4854	9.0	5319	0.1129	0.1129	0.1887	-0.0325	0.0155
0.3118	10.0	5910	0.1187	0.1187	0.2019	-0.0854	0.0578

athirorg/USS-reward-model-utterance-with-history

USS-reward-model-utterance-with-history