From 12597edea67e785919e4ab3dd7851780ac4bddec Mon Sep 17 00:00:00 2001
From: Flaick <471570879@qq.com>
Date: Fri, 25 Apr 2025 10:09:33 +0000
Subject: [PATCH] hle biomed test

---
 .../datasets/HLE/hle_llmverify_gen_6ff468.py       |  6 +++---
 opencompass/datasets/hle.py                        | 14 ++++++++++++++
 2 files changed, 17 insertions(+), 3 deletions(-)

diff --git a/opencompass/configs/datasets/HLE/hle_llmverify_gen_6ff468.py b/opencompass/configs/datasets/HLE/hle_llmverify_gen_6ff468.py
index bb6f40bf..79e579a3 100644
--- a/opencompass/configs/datasets/HLE/hle_llmverify_gen_6ff468.py
+++ b/opencompass/configs/datasets/HLE/hle_llmverify_gen_6ff468.py
@@ -3,7 +3,7 @@ from opencompass.openicl.icl_retriever import ZeroRetriever
 from opencompass.openicl.icl_inferencer import GenInferencer
 from opencompass.evaluator import GenericLLMEvaluator
 from opencompass.datasets import generic_llmjudge_postprocess
-from opencompass.datasets import HLEDataset
+from opencompass.datasets import HLEDataset_Bio_Med
 
 # ----------------------------- Detailed Config -----------------------------
 
@@ -68,7 +68,7 @@ math_eval_cfg = dict(
             ]),
         ),
         dataset_cfg=dict(
-            type=HLEDataset,
+            type=HLEDataset_Bio_Med,
             path='cais/hle',
             reader_cfg=math_reader_cfg,
         ),
@@ -81,7 +81,7 @@ math_eval_cfg = dict(
 
 hle_datasets = [
     dict(
-        type=HLEDataset,
+        type=HLEDataset_Bio_Med,
         abbr='hle_llmjudge',
         path='cais/hle',
         reader_cfg=math_reader_cfg,
diff --git a/opencompass/datasets/hle.py b/opencompass/datasets/hle.py
index 2d7cf74b..80e56a77 100644
--- a/opencompass/datasets/hle.py
+++ b/opencompass/datasets/hle.py
@@ -15,3 +15,17 @@ class HLEDataset(BaseDataset):
         dataset['test'] = dataset['test'].rename_column('question', 'problem')
         dataset['train'] = dataset['test']
         return dataset
+
+
+@LOAD_DATASET.register_module()
+class HLEDataset_Bio_Med(BaseDataset):
+
+    @staticmethod
+    def load(path: str):
+        dataset = load_dataset(path)
+        dataset['test'] = dataset['test'].filter(lambda x: x['image'] == '')
+        dataset['test'] = dataset['test'].filter(
+            lambda x: x['category'] == 'Biology/Medicine')
+        dataset['test'] = dataset['test'].rename_column('question', 'problem')
+        dataset['train'] = dataset['test']
+        return dataset