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
191.9917	1.0	591	0.6336	0.4487	0.5357	-3.1032	0.0274
148.7267	2.0	1182	0.6656	0.8462	0.7999	-6.7382	0.0146
26.3361	3.0	1773	0.7954	0.5077	0.5306	-3.6425	0.0089
33.5642	4.0	2364	0.6652	0.4441	0.5234	-3.0611	-0.0035
8.4169	5.0	2955	0.5751	0.5384	0.5760	-3.9235	-0.0103
4.1928	6.0	3546	0.5960	0.6119	0.6480	-4.5953	-0.0086
2.9277	7.0	4137	0.6578	0.6077	0.6226	-4.5575	-0.0137
1.6562	8.0	4728	0.6687	0.5556	0.6072	-4.0804	-0.0290
1.0501	9.0	5319	0.6575	0.5106	0.5799	-3.6690	0.0028
0.4940	10.0	5910	0.6613	0.5591	0.6169	-4.1125	-0.0015

athirorg/USS-reward-model-utterance-WRS-3

USS-reward-model-utterance-WRS-3