[Feature] Add LEval datasets

Co-authored-by: kennymckormick <dhd@pku.edu.cn>

configs/datasets/LEvalPaperAssistant/LEval_paper_assistant_gen.py

+from mmengine.config import read_base
+
+with read_base():
+    from .LEval_paper_assistant_gen_6c03d0 import LEval_ps_summ_datasets  # noqa: F401, F403

configs/datasets/LEvalPaperAssistant/LEval_paper_assistant_gen_6c03d0.py

+from opencompass.openicl.icl_prompt_template import PromptTemplate
+from opencompass.openicl.icl_retriever import ZeroRetriever
+from opencompass.openicl.icl_inferencer import GenInferencer
+from opencompass.openicl.icl_evaluator import EMEvaluator, RougeEvaluator, SquadEvaluator
+from opencompass.datasets import LEvalPaperAssistantDataset
+
+LEval_ps_summ_reader_cfg = dict(
+    input_columns=['context', 'question'],
+    output_column='answer',
+    train_split='test',
+    test_split='test'
+)
+
+LEval_ps_summ_infer_cfg = dict(
+    prompt_template=dict(
+        type=PromptTemplate,
+        template=dict(
+            round=[
+                dict(role='HUMAN', prompt='{context}\nQuestion: {question}\nAnswer:'),
+                dict(role='BOT', prompt=''),
+            ], )),
+    retriever=dict(type=ZeroRetriever),
+    inferencer=dict(type=GenInferencer, max_out_len=512)
+)
+
+LEval_ps_summ_eval_cfg = dict(
+    evaluator=dict(type=RougeEvaluator), 
+    pred_role='BOT'
+)
+
+LEval_ps_summ_datasets = [
+    dict(
+        type=LEvalPaperAssistantDataset,
+        abbr='LEval_paper_assistant',
+        path='L4NLP/LEval',
+        name='paper_assistant',
+        reader_cfg=LEval_ps_summ_reader_cfg,
+        infer_cfg=LEval_ps_summ_infer_cfg,
+        eval_cfg=LEval_ps_summ_eval_cfg)
+]

configs/datasets/LEvalPatentSumm/LEval_patent_summ_gen.py

+from mmengine.config import read_base
+
+with read_base():
+    from .LEval_patent_summ_gen_db3565 import LEval_patent_summ_datasets  # noqa: F401, F403

configs/datasets/LEvalPatentSumm/LEval_patent_summ_gen_db3565.py

+from opencompass.openicl.icl_prompt_template import PromptTemplate
+from opencompass.openicl.icl_retriever import ZeroRetriever
+from opencompass.openicl.icl_inferencer import GenInferencer
+from opencompass.openicl.icl_evaluator import EMEvaluator, RougeEvaluator, SquadEvaluator
+from opencompass.datasets import LEvalPatentSummDataset
+
+LEval_patent_summ_reader_cfg = dict(
+    input_columns=['context', 'question'],
+    output_column='answer',
+    train_split='test',
+    test_split='test'
+)
+
+LEval_patent_summ_infer_cfg = dict(
+    prompt_template=dict(
+        type=PromptTemplate,
+        template=dict(
+            round=[
+                dict(role='HUMAN', prompt='{context}\n{question}\nTL;DR:'),
+                dict(role='BOT', prompt=''),
+            ], )),
+    retriever=dict(type=ZeroRetriever),
+    inferencer=dict(type=GenInferencer, max_out_len=512)
+)
+
+LEval_patent_summ_eval_cfg = dict(
+    evaluator=dict(type=RougeEvaluator), 
+    pred_role='BOT'
+)
+
+LEval_patent_summ_datasets = [
+    dict(
+        type=LEvalPatentSummDataset,
+        abbr='LEval_patent_summ',
+        path='L4NLP/LEval',
+        name='patent_summ',
+        reader_cfg=LEval_patent_summ_reader_cfg,
+        infer_cfg=LEval_patent_summ_infer_cfg,
+        eval_cfg=LEval_patent_summ_eval_cfg)
+]

configs/datasets/LEvalQuality/LEval_quality_gen.py

+from mmengine.config import read_base
+
+with read_base():
+    from .LEval_quality_gen_bd35f4 import LEval_quality_datasets  # noqa: F401, F403

configs/datasets/LEvalQuality/LEval_quality_gen_bd35f4.py

+from opencompass.openicl.icl_prompt_template import PromptTemplate
+from opencompass.openicl.icl_retriever import ZeroRetriever
+from opencompass.openicl.icl_inferencer import GenInferencer
+from opencompass.openicl.icl_evaluator import EMEvaluator, RougeEvaluator, SquadEvaluator, AccEvaluator
+from opencompass.datasets import LEvalQualityDataset
+from opencompass.utils.text_postprocessors import first_capital_postprocess, first_capital_postprocess_multi
+
+LEval_quality_reader_cfg = dict(
+    input_columns=['context', 'question'],
+    output_column='answer',
+    train_split='test',
+    test_split='test'
+)
+
+LEval_quality_infer_cfg = dict(
+    prompt_template=dict(
+        type=PromptTemplate,
+        template=dict(
+            round=[
+                dict(role='HUMAN', prompt='{context}\nQuestion: {question}\nAnswer:'),
+                dict(role='BOT', prompt=''),
+            ], )),
+    retriever=dict(type=ZeroRetriever),
+    inferencer=dict(type=GenInferencer, max_out_len=10)
+)
+
+LEval_quality_eval_cfg = dict(
+    evaluator=dict(type=AccEvaluator), 
+    pred_postprocessor=dict(type=first_capital_postprocess),
+    pred_role='BOT'
+)
+
+LEval_quality_datasets = [
+    dict(
+        type=LEvalQualityDataset,
+        abbr='LEval_quality',
+        path='L4NLP/LEval',
+        name='quality',
+        reader_cfg=LEval_quality_reader_cfg,
+        infer_cfg=LEval_quality_infer_cfg,
+        eval_cfg=LEval_quality_eval_cfg)
+]

configs/datasets/LEvalReviewSumm/LEval_review_summ_gen.py

+from mmengine.config import read_base
+
+with read_base():
+    from .LEval_review_summ_gen_6c03d0 import LEval_review_summ_datasets  # noqa: F401, F403

configs/datasets/LEvalReviewSumm/LEval_review_summ_gen_6c03d0.py

+from opencompass.openicl.icl_prompt_template import PromptTemplate
+from opencompass.openicl.icl_retriever import ZeroRetriever
+from opencompass.openicl.icl_inferencer import GenInferencer
+from opencompass.openicl.icl_evaluator import EMEvaluator, RougeEvaluator, SquadEvaluator
+from opencompass.datasets import LEvalReviewSummDataset
+
+LEval_review_summ_reader_cfg = dict(
+    input_columns=['context', 'question'],
+    output_column='answer',
+    train_split='test',
+    test_split='test'
+)
+
+LEval_review_summ_infer_cfg = dict(
+    prompt_template=dict(
+        type=PromptTemplate,
+        template=dict(
+            round=[
+                dict(role='HUMAN', prompt='{context}\nQuestion: {question}\nAnswer:'),
+                dict(role='BOT', prompt=''),
+            ], )),
+    retriever=dict(type=ZeroRetriever),
+    inferencer=dict(type=GenInferencer, max_out_len=512)
+)
+
+LEval_review_summ_eval_cfg = dict(
+    evaluator=dict(type=RougeEvaluator), 
+    pred_role='BOT'
+)
+
+LEval_review_summ_datasets = [
+    dict(
+        type=LEvalReviewSummDataset,
+        abbr='LEval_review_summ',
+        path='L4NLP/LEval',
+        name='review_summ',
+        reader_cfg=LEval_review_summ_reader_cfg,
+        infer_cfg=LEval_review_summ_infer_cfg,
+        eval_cfg=LEval_review_summ_eval_cfg)
+]

configs/datasets/LEvalScientificQA/LEval_scientificqa_gen.py

+from mmengine.config import read_base
+
+with read_base():
+    from .LEval_scientificqa_gen_0c6e71 import LEval_scientificqa_datasets  # noqa: F401, F403

configs/datasets/LEvalScientificQA/LEval_scientificqa_gen_0c6e71.py

+from opencompass.openicl.icl_prompt_template import PromptTemplate
+from opencompass.openicl.icl_retriever import ZeroRetriever
+from opencompass.openicl.icl_inferencer import GenInferencer
+from opencompass.openicl.icl_evaluator import EMEvaluator, RougeEvaluator
+from opencompass.datasets import LEvalScientificQADataset
+
+LEval_scientificqa_reader_cfg = dict(
+    input_columns=['context', 'question'],
+    output_column='answer',
+    train_split='test',
+    test_split='test'
+)
+
+LEval_scientificqa_infer_cfg = dict(
+    prompt_template=dict(
+        type=PromptTemplate,
+        template=dict(
+            round=[
+                dict(role='HUMAN', prompt='{context}\nQuestion: {question}\nAnswer:'),
+                dict(role='BOT', prompt=''),
+            ], )),
+    retriever=dict(type=ZeroRetriever),
+    inferencer=dict(type=GenInferencer, max_out_len=64)
+)
+
+LEval_scientificqa_eval_cfg = dict(
+    evaluator=dict(type=RougeEvaluator), 
+    pred_role='BOT'
+)
+
+LEval_scientificqa_datasets = [
+    dict(
+        type=LEvalScientificQADataset,
+        abbr='LEval_scientificqa',
+        path='L4NLP/LEval',
+        name='scientific_qa',
+        reader_cfg=LEval_scientificqa_reader_cfg,
+        infer_cfg=LEval_scientificqa_infer_cfg,
+        eval_cfg=LEval_scientificqa_eval_cfg)
+]

configs/datasets/LEvalTPO/LEval_tpo_gen.py

+from mmengine.config import read_base
+
+with read_base():
+    from .LEval_tpo_gen_bd35f4 import LEval_tpo_datasets

configs/datasets/LEvalTPO/LEval_tpo_gen_bd35f4.py

+from opencompass.openicl.icl_prompt_template import PromptTemplate
+from opencompass.openicl.icl_retriever import ZeroRetriever
+from opencompass.openicl.icl_inferencer import GenInferencer
+from opencompass.openicl.icl_evaluator import EMEvaluator, RougeEvaluator, SquadEvaluator, AccEvaluator
+from opencompass.datasets import LEvalTPODataset
+from opencompass.utils.text_postprocessors import first_capital_postprocess, first_capital_postprocess_multi
+
+LEval_tpo_reader_cfg = dict(
+    input_columns=['context', 'question'],
+    output_column='answer',
+    train_split='test',
+    test_split='test'
+)
+
+LEval_tpo_infer_cfg = dict(
+    prompt_template=dict(
+        type=PromptTemplate,
+        template=dict(
+            round=[
+                dict(role='HUMAN', prompt='{context}\nQuestion: {question}\nAnswer:'),
+                dict(role='BOT', prompt=''),
+            ], )),
+    retriever=dict(type=ZeroRetriever),
+    inferencer=dict(type=GenInferencer, max_out_len=10)
+)
+
+LEval_tpo_eval_cfg = dict(
+    evaluator=dict(type=AccEvaluator), 
+    pred_postprocessor=dict(type=first_capital_postprocess),
+    pred_role='BOT'
+)
+
+LEval_tpo_datasets = [
+    dict(
+        type=LEvalTPODataset,
+        abbr='LEval_tpo',
+        path='L4NLP/LEval',
+        name='tpo',
+        reader_cfg=LEval_tpo_reader_cfg,
+        infer_cfg=LEval_tpo_infer_cfg,
+        eval_cfg=LEval_tpo_eval_cfg)
+]

configs/datasets/LEvalTVShowSumm/.ipynb_checkpoints/LEval_tvshow_summ_gen-checkpoint.py

+from mmengine.config import read_base
+
+with read_base():
+    from .LEval_tvshow_summ_gen_rouge import LEval_tvshow_summ_datasets  # noqa: F401, F403

configs/datasets/LEvalTVShowSumm/.ipynb_checkpoints/LEval_tvshow_summ_gen_rouge-checkpoint.py

+from opencompass.openicl.icl_prompt_template import PromptTemplate
+from opencompass.openicl.icl_retriever import ZeroRetriever
+from opencompass.openicl.icl_inferencer import GenInferencer
+from opencompass.openicl.icl_evaluator import EMEvaluator, RougeEvaluator, SquadEvaluator
+from opencompass.datasets import LEvalTVShowSummDataset
+
+LEval_tvshow_summ_reader_cfg = dict(
+    input_columns=['context', 'question'],
+    output_column='answer',
+    train_split='test',
+    test_split='test'
+)
+
+LEval_tvshow_summ_infer_cfg = dict(
+    prompt_template=dict(
+        type=PromptTemplate,
+        template=dict(
+            round=[
+                dict(role='HUMAN', prompt='{context}\nQuestion: {question}'),
+                dict(role='BOT', prompt='TL;DR:'),
+            ], )),
+    retriever=dict(type=ZeroRetriever),
+    inferencer=dict(type=GenInferencer, max_out_len=512)
+)
+
+LEval_tvshow_summ_eval_cfg = dict(
+    evaluator=dict(type=RougeEvaluator), 
+    pred_role='BOT'
+)
+
+LEval_tvshow_summ_datasets = [
+    dict(
+        type=LEvalTVShowSummDataset,
+        abbr='LEval_tvshow_summ',
+        path='L4NLP/LEval',
+        name='tv_show_summ',
+        reader_cfg=LEval_tvshow_summ_reader_cfg,
+        infer_cfg=LEval_tvshow_summ_infer_cfg,
+        eval_cfg=LEval_tvshow_summ_eval_cfg)
+]

configs/datasets/LEvalTVShowSumm/LEval_tvshow_summ_gen.py

+from mmengine.config import read_base
+
+with read_base():
+    from .LEval_tvshow_summ_gen_049a5c import LEval_tvshow_summ_datasets  # noqa: F401, F403

configs/datasets/LEvalTVShowSumm/LEval_tvshow_summ_gen_049a5c.py

+from opencompass.openicl.icl_prompt_template import PromptTemplate
+from opencompass.openicl.icl_retriever import ZeroRetriever
+from opencompass.openicl.icl_inferencer import GenInferencer
+from opencompass.openicl.icl_evaluator import EMEvaluator, RougeEvaluator, SquadEvaluator
+from opencompass.datasets import LEvalTVShowSummDataset
+
+LEval_tvshow_summ_reader_cfg = dict(
+    input_columns=['context', 'question'],
+    output_column='answer',
+    train_split='test',
+    test_split='test'
+)
+
+LEval_tvshow_summ_infer_cfg = dict(
+    prompt_template=dict(
+        type=PromptTemplate,
+        template=dict(
+            round=[
+                dict(role='HUMAN', prompt='{context}\nQuestion: {question}\nTL;DR:'),
+                dict(role='BOT', prompt=''),
+            ], )),
+    retriever=dict(type=ZeroRetriever),
+    inferencer=dict(type=GenInferencer, max_out_len=512)
+)
+
+LEval_tvshow_summ_eval_cfg = dict(
+    evaluator=dict(type=RougeEvaluator), 
+    pred_role='BOT'
+)
+
+LEval_tvshow_summ_datasets = [
+    dict(
+        type=LEvalTVShowSummDataset,
+        abbr='LEval_tvshow_summ',
+        path='L4NLP/LEval',
+        name='tv_show_summ',
+        reader_cfg=LEval_tvshow_summ_reader_cfg,
+        infer_cfg=LEval_tvshow_summ_infer_cfg,
+        eval_cfg=LEval_tvshow_summ_eval_cfg)
+]

configs/datasets/LEvalTopicRetrieval/LEval_topic_retrieval_gen.py

+from mmengine.config import read_base
+
+with read_base():
+    from .LEval_topic_retrieval_gen_af0562 import LEval_tr_datasets

configs/datasets/LEvalTopicRetrieval/LEval_topic_retrieval_gen_af0562.py

+from opencompass.openicl.icl_prompt_template import PromptTemplate
+from opencompass.openicl.icl_retriever import ZeroRetriever
+from opencompass.openicl.icl_inferencer import GenInferencer
+from opencompass.openicl.icl_evaluator import EMEvaluator, RougeEvaluator, SquadEvaluator, AccEvaluator
+from opencompass.datasets import LEvalTopicRetrievalDataset
+from opencompass.utils.text_postprocessors import first_capital_postprocess, first_capital_postprocess_multi, general_postprocess
+
+LEval_tr_reader_cfg = dict(
+    input_columns=['context', 'question'],
+    output_column='answer',
+    train_split='test',
+    test_split='test'
+)
+
+LEval_tr_infer_cfg = dict(
+    prompt_template=dict(
+        type=PromptTemplate,
+        template=dict(
+            round=[
+                dict(role='HUMAN', prompt='{context}\nQuestion: {question}\nAnswer:'),
+                dict(role='BOT', prompt=''),
+            ], )),
+    retriever=dict(type=ZeroRetriever),
+    inferencer=dict(type=GenInferencer, max_out_len=30)
+)
+
+LEval_tr_eval_cfg = dict(
+    evaluator=dict(type=EMEvaluator), 
+    pred_postprocessor=dict(type=general_postprocess),
+    pred_role='BOT'
+)
+
+LEval_tr_datasets = [
+    dict(
+        type=LEvalTopicRetrievalDataset,
+        abbr='LEval_topic_retrieval',
+        path='L4NLP/LEval',
+        name='topic_retrieval_longchat',
+        reader_cfg=LEval_tr_reader_cfg,
+        infer_cfg=LEval_tr_infer_cfg,
+        eval_cfg=LEval_tr_eval_cfg)
+]

configs/datasets/agieval/agieval_gen.py

 from mmengine.config import read_base
 
 with read_base():
-    from .agieval_gen_397d81 import agieval_datasets  # noqa: F401, F403
+    from .agieval_gen_64afd3 import agieval_datasets  # noqa: F401, F403