Update MathBench (#1176)

* Add Math Evaluation with Judge Model Evaluator

* Add Math Evaluation with Judge Model Evaluator

* Add Math Evaluation with Judge Model Evaluator

* Add Math Evaluation with Judge Model Evaluator

* Fix Llama-3 meta template

* Fix MATH with JudgeLM Evaluation

* Fix MATH with JudgeLM Evaluation

* Fix MATH with JudgeLM Evaluation

* Fix MATH with JudgeLM Evaluation

* Update acclerator

* Update MathBench

---------
Co-authored-by: liuhongwei <liuhongwei@pjlab.org.cn>

configs/datasets/MathBench/mathbench_2024_gen_649349.py → configs/datasets/MathBench/mathbench_2024_gen_19e486.py

@@ -91,7 +91,7 @@ for _split in list(mathbench_sets.keys()):
                     ),
             ),
             retriever=dict(type=ZeroRetriever),
-            inferencer=dict(type=GenInferencer, max_out_len=512),
+            inferencer=dict(type=GenInferencer, max_out_len=2048),
         )
 
         mathbench_eval_cfg = dict(

configs/datasets/MathBench/mathbench_gen.py

 from mmengine.config import read_base
 
 with read_base():
-    from .mathbench_2024_gen_649349 import mathbench_datasets  # noqa: F401, F403
+    from .mathbench_2024_gen_19e486 import mathbench_datasets  # noqa: F401, F403