[Feature] Support eval WildBench-Score

examples/eval_OpenHuEval_HuLifeQA.py

@@ -17,11 +17,10 @@ with read_base():
 
     from opencompass.configs.models.hf_internlm.lmdeploy_internlm3_8b_instruct import models as lmdeploy_internlm3_8b_instruct_model
 
-    from opencompass.configs.models.openai.o1_mini_2024_09_12 import models as o1_mini_2024_09_12_model
     from opencompass.configs.models.qwq.lmdeploy_qwq_32b_preview import models as lmdeploy_qwq_32b_preview_model
     from opencompass.configs.models.deepseek.deepseek_r1_api_aliyun import models as deepseek_r1_api_aliyun_model
-    from opencompass.configs.models.deepseek.deepseek_r1_distill_llama_8b_api_aliyun import models as deepseek_r1_distill_llama_8b_api_aliyun_model
+    from opencompass.configs.models.openai.o1_mini_2024_09_12 import models as o1_mini_2024_09_12_model
-    from opencompass.configs.models.deepseek.deepseek_r1_distill_qwen_7b_api_aliyun import models as deepseek_r1_distill_qwen_7b_api_aliyun_model
+    # from opencompass.configs.models.openai.o3_mini_2025_01_31 import models as o3_mini_2025_01_31_model
 
 from opencompass.models import OpenAI
 from opencompass.partitioners import (
@@ -46,12 +45,6 @@ for model in deepseek_r1_api_aliyun_model:
             'type': 'rm_<think>_before_eval'
         }
     }
-    if model['abbr'].startswith('QwQ'):
-        model['pred_postprocessor'] = {
-            'OpenHuEval_*': {
-                'type': 'extract_qwq_answer_before_eval'
-            }
-        }
 del model
 
 models = [
@@ -60,8 +53,6 @@ models = [
     *o1_mini_2024_09_12_model,
     *deepseek_v3_api_aliyun_model,
     *deepseek_r1_api_aliyun_model,
-    *deepseek_r1_distill_llama_8b_api_aliyun_model,
-    *deepseek_r1_distill_qwen_7b_api_aliyun_model,
     *lmdeploy_qwen2_5_7b_instruct_model,
     *lmdeploy_qwen2_5_72b_instruct_model,
     *lmdeploy_llama3_1_8b_instruct_model,

examples/eval_subjective_wildbench_single.py

@@ -0,0 +1,112 @@
+from mmengine.config import read_base
+
+with read_base():
+    from opencompass.configs.datasets.subjective.wildbench.wildbench_single_judge import wildbench_datasets
+
+    from opencompass.configs.models.openai.gpt_4o_mini_20240718 import models as gpt_4o_mini_20240718_model
+    from opencompass.configs.models.openai.gpt_4o_2024_11_20 import models as gpt_4o_20241120_model
+    from opencompass.configs.models.deepseek.deepseek_v3_api_aliyun import models as deepseek_v3_api_aliyun_model
+
+    from opencompass.configs.models.qwen2_5.lmdeploy_qwen2_5_7b_instruct import models as lmdeploy_qwen2_5_7b_instruct_model
+    from opencompass.configs.models.qwen2_5.lmdeploy_qwen2_5_72b_instruct import models as lmdeploy_qwen2_5_72b_instruct_model
+    from opencompass.configs.models.hf_llama.lmdeploy_llama3_1_8b_instruct import models as lmdeploy_llama3_1_8b_instruct_model
+    from opencompass.configs.models.hf_llama.lmdeploy_llama3_1_70b_instruct import models as lmdeploy_llama3_1_70b_instruct_model
+
+    # from opencompass.configs.models.hf_internlm.lmdeploy_internlm3_8b_instruct import models as lmdeploy_internlm3_8b_instruct_model
+
+    from opencompass.configs.models.openai.o1_mini_2024_09_12 import models as o1_mini_2024_09_12_model
+    from opencompass.configs.models.qwq.lmdeploy_qwq_32b_preview import models as lmdeploy_qwq_32b_preview_model
+    from opencompass.configs.models.deepseek.deepseek_r1_api_aliyun import models as deepseek_r1_api_aliyun_model
+    # from opencompass.configs.models.openai.o3_mini_2025_01_31 import models as o3_mini_2025_01_31_model
+
+from opencompass.models import OpenAI
+from opencompass.partitioners import (
+    NumWorkerPartitioner,
+    SubjectiveNumWorkerPartitioner,
+)
+from opencompass.runners import LocalRunner, SlurmSequentialRunner
+from opencompass.summarizers import WildBenchSingleSummarizer
+from opencompass.tasks import OpenICLInferTask
+from opencompass.tasks.subjective_eval import SubjectiveEvalTask
+
+api_meta_template = dict(round=[
+    dict(role='SYSTEM', api_role='SYSTEM'),
+    dict(role='HUMAN', api_role='HUMAN'),
+    dict(role='BOT', api_role='BOT', generate=True),
+])
+
+for model in deepseek_r1_api_aliyun_model:
+    model['return_reasoning_content'] = True
+    model['pred_postprocessor'] = {
+        'open_hu_eval_*': {
+            'type': 'rm_<think>_before_eval'
+        }
+    }
+del model
+
+models = [
+    *gpt_4o_mini_20240718_model,
+    *gpt_4o_20241120_model,
+    *o1_mini_2024_09_12_model,
+    *deepseek_v3_api_aliyun_model,
+    *deepseek_r1_api_aliyun_model,
+    *lmdeploy_qwen2_5_7b_instruct_model,
+    *lmdeploy_qwen2_5_72b_instruct_model,
+    *lmdeploy_llama3_1_8b_instruct_model,
+    *lmdeploy_llama3_1_70b_instruct_model,
+    # *lmdeploy_internlm3_8b_instruct_model,
+    *lmdeploy_qwq_32b_preview_model,
+]
+
+judge_models = [
+    dict(
+        abbr='GPT-4o-2024-08-06',
+        type=OpenAI,
+        path='gpt-4o-2024-08-06',
+        key='ENV',
+        openai_proxy_url='ENV',
+        verbose=True,
+        meta_template=api_meta_template,
+        query_per_second=2,
+        max_out_len=8192,
+        max_seq_len=16384,
+        batch_size=8,
+        temperature=0,
+    )
+]
+
+for ds in wildbench_datasets:
+    ds.update(dict(mode='singlescore'))
+del ds
+datasets = [*wildbench_datasets]
+del wildbench_datasets
+
+infer = dict(
+    partitioner=dict(
+        type=NumWorkerPartitioner,
+        num_worker=8,
+    ),
+    runner=dict(
+        type=SlurmSequentialRunner,
+        max_num_workers=16,
+        task=dict(type=OpenICLInferTask),
+    ),
+)
+
+eval = dict(
+    partitioner=dict(
+        type=SubjectiveNumWorkerPartitioner,
+        num_worker=8,
+        models=models,
+        judge_models=judge_models,
+    ),
+    runner=dict(type=LocalRunner,
+                max_num_workers=16,
+                task=dict(type=SubjectiveEvalTask)),
+)
+
+summarizer = dict(type=WildBenchSingleSummarizer)
+
+work_dir = (
+    './outputs/' + __file__.split('/')[-1].split('.')[0] + '/'
+)  # do NOT modify this line, yapf: disable, pylint: disable

opencompass/configs/datasets/subjective/wildbench/wildbench_single_judge.py

@@ -0,0 +1,50 @@
+from opencompass.datasets import WildBenchDataset
+from opencompass.openicl.icl_evaluator import LMEvaluator
+from opencompass.openicl.icl_inferencer import ChatInferencer
+from opencompass.openicl.icl_prompt_template import PromptTemplate
+from opencompass.openicl.icl_retriever import ZeroRetriever
+
+wildbench_reader_cfg = dict(
+    input_columns=['dialogue', 'prompt'],
+    output_column='judge',
+)
+
+data_path ='/mnt/hwfile/opendatalab/yanghaote/datasets/WildBench/wildbench.jsonl'
+
+wildbench_datasets = []
+wildbench_infer_cfg = dict(
+    prompt_template=dict(
+        type=PromptTemplate,
+        template="""{dialogue}"""
+    ),
+    retriever=dict(type=ZeroRetriever),
+    inferencer=dict(
+        type=ChatInferencer,
+        max_seq_len=32768,
+        max_out_len=8192,
+        infer_mode='last',
+    ),
+)
+
+wildbench_eval_cfg = dict(
+    evaluator=dict(
+        type=LMEvaluator,
+        prompt_template=dict(
+            type=PromptTemplate, 
+            template="""{prompt}"""
+        ),
+    ),
+    pred_role='BOT',
+)
+
+wildbench_datasets.append(
+    dict(
+        abbr='wildbench',
+        type=WildBenchDataset,
+        path=data_path,
+        eval_mode='single',
+        reader_cfg=wildbench_reader_cfg,
+        infer_cfg=wildbench_infer_cfg,
+        eval_cfg=wildbench_eval_cfg,
+    )
+)

opencompass/configs/models/openai/gpt_4o_2024_11_20.py

@@ -15,7 +15,7 @@ models = [
         openai_proxy_url='ENV',
         meta_template=api_meta_template,
         query_per_second=1,
-        max_out_len=2048,
+        max_out_len=8192,
-        max_seq_len=4096,
+        max_seq_len=8192,
         batch_size=8),
 ]

opencompass/configs/models/openai/gpt_4o_mini_20240718.py

@@ -15,7 +15,7 @@ models = [
         openai_proxy_url='ENV',
         meta_template=api_meta_template,
         query_per_second=1,
-        max_out_len=2048,
+        max_out_len=8192,
-        max_seq_len=4096,
+        max_seq_len=8192,
         batch_size=8),
 ]