remove priority from evaluators

folding/registries/evaluation_registry.py

@@ -13,9 +13,8 @@
 
 
 class SyntheticMDEvaluator(BaseEvaluator):
-    def __init__(self, pdb_id: str, priority: float = 1, **kwargs):
+    def __init__(self, pdb_id: str, **kwargs):
         self.pdb_id = pdb_id
-        self.priority = priority
         self.kwargs = kwargs
         self.md_simulator = OpenMMSimulation()
 
@@ -316,8 +315,7 @@ def name(self) -> str:
 
 
 class OrganicMDEvaluator(BaseEvaluator):
-    def __init__(self, priority: float = 1, **kwargs):
-        self.priority = priority
+    def __init__(self, **kwargs):
         self.kwargs = kwargs
 
     def _evaluate(self, data: Dict[str, Any]) -> float:
@@ -328,8 +326,7 @@ def name(self) -> str:
 
 
 class SyntheticMLEvaluator(BaseEvaluator):
-    def __init__(self, priority: float = 1, **kwargs):
-        self.priority = priority
+    def __init__(self, **kwargs):
         self.kwargs = kwargs
 
     def _evaluate(self, data: Dict[str, Any]) -> float:
@@ -340,8 +337,7 @@ def name(self) -> str:
 
 
 class OrganicMLEvaluator(BaseEvaluator):
-    def __init__(self, priority: float = 1, **kwargs):
-        self.priority = priority
+    def __init__(self, **kwargs):
         self.kwargs = kwargs
 
     def _evaluate(self, data: Dict[str, Any]) -> float:

folding/rewards/reward_pipeline.py

@@ -1,13 +1,10 @@
 import torch
-import bittensor as bt
 from folding.store import Job
 from folding.rewards.linear_reward import divide_decreasing
 from folding.utils.logger import logger
 
 
-async def reward_pipeline(
-    energies: torch.Tensor, rewards: torch.Tensor, top_reward: float, job: Job
-):
+async def reward_pipeline(energies: torch.Tensor, rewards: torch.Tensor, top_reward: float, job: Job):
     """A reward pipeline that determines how to place rewards onto the miners sampled within the batch.
     Currently applies a linearly decreasing reward on all miners that are not the current best / previously
     best loss using the function "divide_decreasing".
@@ -22,9 +19,7 @@ async def reward_pipeline(
 
     # If the best hotkey is not in the set of hotkeys in the job, this means that the top miner has stopped replying.
     if job.best_hotkey not in job.hotkeys:
-        logger.warning(
-            f"Best hotkey {job.best_hotkey} not in hotkeys {job.hotkeys}. Assigning no reward."
-        )
+        logger.warning(f"Best hotkey {job.best_hotkey} not in hotkeys {job.hotkeys}. Assigning no reward.")
         return rewards  # rewards of all 0s.
 
     best_index = job.hotkeys.index(job.best_hotkey)