transform changes + config updates

Author: Vidhisha Balachandran

eureka_ml_insights/data_utils/transform.py

@@ -505,7 +505,8 @@ def _extract_usage(self, row, usage_completion_read_col):
         Returns:
             int: The token usage for the row.
         """
-        if row[self.prepend_completion_read_col + "is_valid"]:
+        #if row[self.prepend_completion_read_col + "is_valid"]:
+        if row[self.prepend_completion_read_col + "usage"] :
             return row[self.prepend_completion_read_col + "usage"][usage_completion_read_col]
         return np.nan
 

eureka_ml_insights/user_configs/omni_math.py

@@ -69,8 +69,8 @@ def configure_pipeline(self, model_config=None, resume_from=None, eval_resume_fr
             ),
             output_dir=os.path.join(self.log_dir, "inference_result"),
             resume_from=resume_from,
-            max_concurrent=5,
-            #requests_per_minute=5
+            max_concurrent=1,
+            requests_per_minute=5
         )
 
         # eval data preprocessing