change the name to slicpairpm

The training and use of the models are similar to that of Slic paper SLiC-HF: Sequence Likelihood Calibration with Human Feedback.

rewardbench/models/__init__.py

@@ -35,7 +35,7 @@
     build_starling_rm,
 )
 from .ziya import ZiyaPipeline
-from .pair_pm import PairPMPipeline
+from .slicpairpm import SlicPairPMPipeline
 
 # Please open a PR if you need to add more custom modeling code / utilize existing code for you model
 REWARD_MODEL_CONFIG = {

rewardbench/models/pairpm.py → rewardbench/models/slicpairpm.py

@@ -9,7 +9,7 @@
 from typing import List, Optional, Tuple, Union
 
 
-class PairPMPipeline:
+class SlicPairPMPipeline:
 
     def __init__(self, model_path):
         self.model = AutoModelForCausalLM.from_pretrained(model_path,).cuda() #, attn_implementation="flash_attention_2",  torch_dtype=torch.bfloat16