From 07a6dacf33141fdd176c5870574cbba5b73c27e3 Mon Sep 17 00:00:00 2001
From: bittersweet1999 <148421775+bittersweet1999@users.noreply.github.com>
Date: Fri, 24 May 2024 23:30:01 +0800
Subject: [PATCH] fix length (#1180)

---
 .../{arena_hard_scoring.py => arena_hard_compare.py}        | 0
 configs/eval_subjective_arena_hard.py                       | 6 +++---
 2 files changed, 3 insertions(+), 3 deletions(-)
 rename configs/datasets/subjective/arena_hard/{arena_hard_scoring.py => arena_hard_compare.py} (100%)

diff --git a/configs/datasets/subjective/arena_hard/arena_hard_scoring.py b/configs/datasets/subjective/arena_hard/arena_hard_compare.py
similarity index 100%
rename from configs/datasets/subjective/arena_hard/arena_hard_scoring.py
rename to configs/datasets/subjective/arena_hard/arena_hard_compare.py
diff --git a/configs/eval_subjective_arena_hard.py b/configs/eval_subjective_arena_hard.py
index a64d1b76..08db8b29 100644
--- a/configs/eval_subjective_arena_hard.py
+++ b/configs/eval_subjective_arena_hard.py
@@ -14,7 +14,7 @@ from opencompass.tasks.subjective_eval import SubjectiveEvalTask
 from opencompass.summarizers import ArenaHardSummarizer
 
 with read_base():
-    from .datasets.subjective.arena_hard.arena_hard_scoring import subjective_datasets
+    from .datasets.subjective.arena_hard.arena_hard_compare import subjective_datasets
 
 api_meta_template = dict(
     round=[
@@ -72,8 +72,8 @@ judge_models = [dict(
     key='',
         meta_template=api_meta_template,
         query_per_second=1,
-        max_out_len=1024,
-        max_seq_len=4096,
+        max_out_len=5120,
+        max_seq_len=9216,
         batch_size=10,
         retry=10,
         temperature = 0,