debug

src/backend/run_eval_suite_harness.py

@@ -13,6 +13,17 @@ from src.logging import setup_logger
 logging.getLogger("openai").setLevel(logging.WARNING)
 logger = setup_logger(__name__)
 
+class NumpyArrayEncoder(json.JSONEncoder):
+    def default(self, obj):
+        if isinstance(obj, np.ndarray):
+            return obj.tolist()
+        elif isinstance(obj, np.integer):
+            return int(obj)
+        elif isinstance(obj, np.floating):
+            return float(obj)
+        else:
+            return super().default(obj)
+
 def run_evaluation(eval_request: EvalRequest, task_names: list, num_fewshot: int, batch_size: int, device: str, local_dir: str, results_repo: str, no_cache: bool =True, limit: int =None):
     """Runs one evaluation for the current evaluation request file, then pushes the results to the hub.
 
@@ -75,7 +86,7 @@ def run_evaluation(eval_request: EvalRequest, task_names: list, num_fewshot: int
                 print(key2)
                 print(value2)
                 print(type(value))
-    dumped = json.dumps(results, indent=2)
+    dumped = json.dumps(results, cls=NumpyArrayEncoder, indent=2)
     logger.info(dumped)
 
     output_path = os.path.join(local_dir, *eval_request.model.split("/"), f"results_{datetime.now()}.json")