Skip to content

GitLab

Unverified Commit aa2dd2b5 authored by Fengzhe Zhou's avatar Fengzhe Zhou Committed by GitHub
Browse files

[Format] Add config lints (#892)

parent 3dbba119
Expand all Show whitespace changes
Inline Side-by-side
Showing with 613 additions and 613 deletions
configs/datasets/FewCLUE_chid/FewCLUE_chid_ppl_acccb5.py
View file @ aa2dd2b5
......@@ -10,7 +10,7 @@ chid_reader_cfg = dict(
chid_infer_cfg = dict(
prompt_template=dict(
type=PromptTemplate,
template={i: f"以下句子是否通顺？\n{{content{i}}}\n这个句子是通顺的。"
template={i: f'以下句子是否通顺？\n{{content{i}}}\n这个句子是通顺的。'
for i in range(7)}),
retriever=dict(type=ZeroRetriever),
inferencer=dict(type=PPLInferencer))
......
configs/datasets/FewCLUE_cluewsc/FewCLUE_cluewsc_gen_c68933.py
View file @ aa2dd2b5
......@@ -6,8 +6,8 @@ from opencompass.datasets import CluewscDataset_V2
from opencompass.utils.text_postprocessors import first_capital_postprocess
cluewsc_reader_cfg = dict(
input_columns=["span1", "span2", "text", "new_text"],
output_column="label",
input_columns=['span1', 'span2', 'text', 'new_text'],
output_column='label',
)
cluewsc_infer_cfg = dict(
......@@ -15,9 +15,9 @@ cluewsc_infer_cfg = dict(
type=PromptTemplate,
template=dict(round=[
dict(
role="HUMAN",
role='HUMAN',
prompt=
"{text}\n此处，“{span2}”是否指代“{span1}“？\nA. 是\nB. 否\n请从”A“，”B“中进行选择。\n答：",
'{text}\n此处，“{span2}”是否指代“{span1}“？\nA. 是\nB. 否\n请从”A“，”B“中进行选择。\n答：',
),
]),
),
......@@ -27,23 +27,23 @@ cluewsc_infer_cfg = dict(
cluewsc_eval_cfg = dict(
evaluator=dict(type=AccEvaluator),
pred_role="BOT",
pred_role='BOT',
pred_postprocessor=dict(type=first_capital_postprocess),
)
cluewsc_datasets = [
dict(
abbr="cluewsc-dev",
abbr='cluewsc-dev',
type=CluewscDataset_V2,
path="./data/FewCLUE/cluewsc/dev_few_all.json",
path='./data/FewCLUE/cluewsc/dev_few_all.json',
reader_cfg=cluewsc_reader_cfg,
infer_cfg=cluewsc_infer_cfg,
eval_cfg=cluewsc_eval_cfg,
),
dict(
abbr="cluewsc-test",
abbr='cluewsc-test',
type=CluewscDataset_V2,
path="./data/FewCLUE/cluewsc/test_public.json",
path='./data/FewCLUE/cluewsc/test_public.json',
reader_cfg=cluewsc_reader_cfg,
infer_cfg=cluewsc_infer_cfg,
eval_cfg=cluewsc_eval_cfg,
......
configs/datasets/FewCLUE_cluewsc/FewCLUE_cluewsc_ppl_12e4e0.py
View file @ aa2dd2b5
......@@ -15,20 +15,20 @@ cluewsc_infer_cfg = dict(
0:
dict(round=[
dict(
role="HUMAN",
role='HUMAN',
prompt=
"{text}\nHere, is the pronoun \"{span2}\" used to mean \"{span1}\"?"
),
dict(role="BOT", prompt="No.")
dict(role='BOT', prompt='No.')
]),
1:
dict(round=[
dict(
role="HUMAN",
role='HUMAN',
prompt=
"{text}\nHere, is the pronoun \"{span2}\" used to mean \"{span1}\"?"
),
dict(role="BOT", prompt="Yes.")
dict(role='BOT', prompt='Yes.')
]),
}),
retriever=dict(type=ZeroRetriever),
......
configs/datasets/FewCLUE_cluewsc/FewCLUE_cluewsc_ppl_868415.py
View file @ aa2dd2b5
......@@ -15,16 +15,16 @@ cluewsc_infer_cfg = dict(
0:
dict(round=[
dict(
role="HUMAN",
prompt="{text}\n此处，代词“{span2}“被用于指代“{span1}“吗?"),
dict(role="BOT", prompt="否")
role='HUMAN',
prompt='{text}\n此处，代词“{span2}“被用于指代“{span1}“吗?'),
dict(role='BOT', prompt='否')
]),
1:
dict(round=[
dict(
role="HUMAN",
prompt="{text}\n此处，代词“{span2}“被用于指代“{span1}“吗?"),
dict(role="BOT", prompt="是")
role='HUMAN',
prompt='{text}\n此处，代词“{span2}“被用于指代“{span1}“吗?'),
dict(role='BOT', prompt='是')
]),
}),
retriever=dict(type=ZeroRetriever),
......
configs/datasets/FewCLUE_csl/FewCLUE_csl_gen_28b223.py
View file @ aa2dd2b5
......@@ -6,8 +6,8 @@ from opencompass.datasets import CslDataset_V2
from opencompass.utils.text_postprocessors import first_capital_postprocess
csl_reader_cfg = dict(
input_columns=["abst", "keywords"],
output_column="label",
input_columns=['abst', 'keywords'],
output_column='label',
)
csl_infer_cfg = dict(
......@@ -15,9 +15,9 @@ csl_infer_cfg = dict(
type=PromptTemplate,
template=dict(round=[
dict(
role="HUMAN",
role='HUMAN',
prompt=
"摘要是对论文内容不加注释和评论的简短陈述，要求扼要地说明研究工作的目的、研究方法和最终结论等。\n关键词是一篇学术论文的核心词汇，一般由一系列名词组成。关键词在全文中应有较高出现频率，且能起到帮助文献检索的作用。\n摘要：{abst}\n关键词：{keywords}\n请问上述关键词是否匹配摘要且符合要求？\nA. 否\nB. 是\n请从”A“，”B“中进行选择。\n答："
'摘要是对论文内容不加注释和评论的简短陈述，要求扼要地说明研究工作的目的、研究方法和最终结论等。\n关键词是一篇学术论文的核心词汇，一般由一系列名词组成。关键词在全文中应有较高出现频率，且能起到帮助文献检索的作用。\n摘要：{abst}\n关键词：{keywords}\n请问上述关键词是否匹配摘要且符合要求？\nA. 否\nB. 是\n请从”A“，”B“中进行选择。\n答：'
)
]),
),
......@@ -27,23 +27,23 @@ csl_infer_cfg = dict(
csl_eval_cfg = dict(
evaluator=dict(type=AccEvaluator),
pred_role="BOT",
pred_role='BOT',
pred_postprocessor=dict(type=first_capital_postprocess),
)
csl_datasets = [
dict(
abbr="csl_dev",
abbr='csl_dev',
type=CslDataset_V2,
path="./data/FewCLUE/csl/dev_few_all.json",
path='./data/FewCLUE/csl/dev_few_all.json',
reader_cfg=csl_reader_cfg,
infer_cfg=csl_infer_cfg,
eval_cfg=csl_eval_cfg,
),
dict(
abbr="csl_test",
abbr='csl_test',
type=CslDataset_V2,
path="./data/FewCLUE/csl/test_public.json",
path='./data/FewCLUE/csl/test_public.json',
reader_cfg=csl_reader_cfg,
infer_cfg=csl_infer_cfg,
eval_cfg=csl_eval_cfg,
......
configs/datasets/FewCLUE_csl/FewCLUE_csl_gen_87f4a8.py
View file @ aa2dd2b5
......@@ -6,8 +6,8 @@ from opencompass.datasets import CslDataset_V2
from opencompass.utils.text_postprocessors import first_capital_postprocess
csl_reader_cfg = dict(
input_columns=["abst", "keywords"],
output_column="label",
input_columns=['abst', 'keywords'],
output_column='label',
)
csl_infer_cfg = dict(
......@@ -15,9 +15,9 @@ csl_infer_cfg = dict(
type=PromptTemplate,
template=dict(round=[
dict(
role="HUMAN",
role='HUMAN',
prompt=
"摘要：{abst}\n关键词：{keywords}\n上述关键词出现在学术期刊中是否恰当？\nA. 否\nB. 是\n请从”A“，”B“中进行选择。\n答："
'摘要：{abst}\n关键词：{keywords}\n上述关键词出现在学术期刊中是否恰当？\nA. 否\nB. 是\n请从”A“，”B“中进行选择。\n答：'
)
]),
),
......@@ -27,23 +27,23 @@ csl_infer_cfg = dict(
csl_eval_cfg = dict(
evaluator=dict(type=AccEvaluator),
pred_role="BOT",
pred_role='BOT',
pred_postprocessor=dict(type=first_capital_postprocess),
)
csl_datasets = [
dict(
abbr="csl_dev",
abbr='csl_dev',
type=CslDataset_V2,
path="./data/FewCLUE/csl/dev_few_all.json",
path='./data/FewCLUE/csl/dev_few_all.json',
reader_cfg=csl_reader_cfg,
infer_cfg=csl_infer_cfg,
eval_cfg=csl_eval_cfg,
),
dict(
abbr="csl_test",
abbr='csl_test',
type=CslDataset_V2,
path="./data/FewCLUE/csl/test_public.json",
path='./data/FewCLUE/csl/test_public.json',
reader_cfg=csl_reader_cfg,
infer_cfg=csl_infer_cfg,
eval_cfg=csl_eval_cfg,
......
configs/datasets/FewCLUE_csl/FewCLUE_csl_ppl_769f8d.py
View file @ aa2dd2b5
......@@ -5,17 +5,17 @@ from opencompass.openicl.icl_evaluator import AccEvaluator
from opencompass.datasets import CslDataset
csl_reader_cfg = dict(
input_columns=["abst", "keywords"], output_column='label')
input_columns=['abst', 'keywords'], output_column='label')
csl_infer_cfg = dict(
prompt_template=dict(
type=PromptTemplate,
template={
0:
dict(round=[dict(role="HUMAN", prompt="摘要：{abst}")]),
dict(round=[dict(role='HUMAN', prompt='摘要：{abst}')]),
1:
dict(
round=[dict(role="HUMAN", prompt="摘要：{abst}\n关键词：{keywords}")
round=[dict(role='HUMAN', prompt='摘要：{abst}\n关键词：{keywords}')
]),
}),
retriever=dict(type=ZeroRetriever),
......
configs/datasets/FewCLUE_csl/FewCLUE_csl_ppl_841b62.py
View file @ aa2dd2b5
......@@ -5,14 +5,14 @@ from opencompass.openicl.icl_evaluator import AccEvaluator
from opencompass.datasets import CslDataset
csl_reader_cfg = dict(
input_columns=["abst", "keywords"], output_column='label')
input_columns=['abst', 'keywords'], output_column='label')
csl_infer_cfg = dict(
prompt_template=dict(
type=PromptTemplate,
template={
0: "摘要：{abst}",
1: "摘要：{abst}\n关键词：{keywords}"
0: '摘要：{abst}',
1: '摘要：{abst}\n关键词：{keywords}'
}),
retriever=dict(type=ZeroRetriever),
inferencer=dict(type=PPLInferencer))
......
configs/datasets/FewCLUE_eprstmt/FewCLUE_eprstmt_gen_740ea0.py
View file @ aa2dd2b5
......@@ -6,14 +6,14 @@ from opencompass.datasets import eprstmtDataset_V2
from opencompass.utils.text_postprocessors import first_capital_postprocess
eprstmt_reader_cfg = dict(
input_columns=["sentence"], output_column="label", test_split="train")
input_columns=['sentence'], output_column='label', test_split='train')
eprstmt_infer_cfg = dict(
prompt_template=dict(
type=PromptTemplate,
template=dict(round=[
dict(
role="HUMAN",
role='HUMAN',
prompt=
'内容： "{sentence}"。请对上述内容进行情绪分类。\nA. 积极\nB. 消极\n请从”A“，”B“中进行选择。\n答：'
),
......@@ -25,23 +25,23 @@ eprstmt_infer_cfg = dict(
eprstmt_eval_cfg = dict(
evaluator=dict(type=AccEvaluator),
pred_role="BOT",
pred_role='BOT',
pred_postprocessor=dict(type=first_capital_postprocess),
)
eprstmt_datasets = [
dict(
abbr="eprstmt-dev",
abbr='eprstmt-dev',
type=eprstmtDataset_V2,
path="./data/FewCLUE/eprstmt/dev_few_all.json",
path='./data/FewCLUE/eprstmt/dev_few_all.json',
reader_cfg=eprstmt_reader_cfg,
infer_cfg=eprstmt_infer_cfg,
eval_cfg=eprstmt_eval_cfg,
),
dict(
abbr="eprstmt-test",
abbr='eprstmt-test',
type=eprstmtDataset_V2,
path="./data/FewCLUE/eprstmt/test_public.json",
path='./data/FewCLUE/eprstmt/test_public.json',
reader_cfg=eprstmt_reader_cfg,
infer_cfg=eprstmt_infer_cfg,
eval_cfg=eprstmt_eval_cfg,
......
configs/datasets/FewCLUE_ocnli_fc/FewCLUE_ocnli_fc_gen_f97a97.py
View file @ aa2dd2b5
......@@ -6,18 +6,18 @@ from opencompass.datasets import cmnliDataset_V2
from opencompass.utils.text_postprocessors import first_capital_postprocess
ocnli_fc_reader_cfg = dict(
input_columns=["sentence1", "sentence2"],
output_column="label",
test_split="train")
input_columns=['sentence1', 'sentence2'],
output_column='label',
test_split='train')
ocnli_fc_infer_cfg = dict(
prompt_template=dict(
type=PromptTemplate,
template=dict(round=[
dict(
role="HUMAN",
role='HUMAN',
prompt=
"阅读文章：{sentence1}\n根据上文，回答如下问题：{sentence2}\nA. 对\nB. 错\nC. 可能\n请从“A”，“B”，“C”中进行选择。\n答："
'阅读文章：{sentence1}\n根据上文，回答如下问题：{sentence2}\nA. 对\nB. 错\nC. 可能\n请从“A”，“B”，“C”中进行选择。\n答：'
),
]),
),
......@@ -26,23 +26,23 @@ ocnli_fc_infer_cfg = dict(
)
ocnli_fc_eval_cfg = dict(
evaluator=dict(type=AccEvaluator),
pred_role="BOT",
pred_role='BOT',
pred_postprocessor=dict(type=first_capital_postprocess),
)
ocnli_fc_datasets = [
dict(
abbr="ocnli_fc-dev",
abbr='ocnli_fc-dev',
type=cmnliDataset_V2, # ocnli_fc share the same format with cmnli
path="./data/FewCLUE/ocnli/dev_few_all.json",
path='./data/FewCLUE/ocnli/dev_few_all.json',
reader_cfg=ocnli_fc_reader_cfg,
infer_cfg=ocnli_fc_infer_cfg,
eval_cfg=ocnli_fc_eval_cfg,
),
dict(
abbr="ocnli_fc-test",
abbr='ocnli_fc-test',
type=cmnliDataset_V2, # ocnli_fc share the same format with cmnli
path="./data/FewCLUE/ocnli/test_public.json",
path='./data/FewCLUE/ocnli/test_public.json',
reader_cfg=ocnli_fc_reader_cfg,
infer_cfg=ocnli_fc_infer_cfg,
eval_cfg=ocnli_fc_eval_cfg,
......
configs/datasets/FewCLUE_ocnli_fc/FewCLUE_ocnli_fc_ppl_9e8b3d.py
View file @ aa2dd2b5
......@@ -16,22 +16,22 @@ ocnli_fc_infer_cfg = dict(
'contradiction':
dict(round=[
dict(
role="HUMAN",
prompt="阅读文章：{sentence1}\n根据上文，回答如下问题：{sentence2}？"),
dict(role="BOT", prompt="错")
role='HUMAN',
prompt='阅读文章：{sentence1}\n根据上文，回答如下问题：{sentence2}？'),
dict(role='BOT', prompt='错')
]),
'entailment':
dict(round=[
dict(
role="HUMAN",
prompt="阅读文章：{sentence1}\n根据上文，回答如下问题：{sentence2}？"),
dict(role="BOT", prompt="对")
role='HUMAN',
prompt='阅读文章：{sentence1}\n根据上文，回答如下问题：{sentence2}？'),
dict(role='BOT', prompt='对')
]),
'neutral':
dict(round=[
dict(
role="HUMAN", prompt="如果{sentence1}为真，那么{sentence2}也为真吗？"),
dict(role="BOT", prompt="可能")
role='HUMAN', prompt='如果{sentence1}为真，那么{sentence2}也为真吗？'),
dict(role='BOT', prompt='可能')
]),
}),
retriever=dict(type=ZeroRetriever),
......
configs/datasets/FewCLUE_tnews/FewCLUE_tnews_gen_b90e4a.py
View file @ aa2dd2b5
......@@ -6,30 +6,30 @@ from opencompass.datasets import TNewsDataset_V2
from opencompass.utils.text_postprocessors import first_capital_postprocess
tnews_reader_cfg = dict(
input_columns="sentence",
output_column="label_desc2",
input_columns='sentence',
output_column='label_desc2',
)
tnews_labels = [
"农业新闻", # news_agriculture
"旅游新闻", # news_travel
"游戏新闻", # news_game
"科技类别公司新闻", # news_tech
"体育类别新闻", # news_sports
"初升高教育新闻", # news_edu
"娱乐圈新闻", # news_entertainment
"投资资讯", # news_finance
"军事类别常识", # news_military
"车辆新闻", # news_car
"楼市新闻", # news_house
"环球不含中国类别新闻", # news_world
"书籍文化历史类别新闻", # news_culture
"故事类别新闻", # news_story
"股票市场类别新闻", # news_stock
'农业新闻', # news_agriculture
'旅游新闻', # news_travel
'游戏新闻', # news_game
'科技类别公司新闻', # news_tech
'体育类别新闻', # news_sports
'初升高教育新闻', # news_edu
'娱乐圈新闻', # news_entertainment
'投资资讯', # news_finance
'军事类别常识', # news_military
'车辆新闻', # news_car
'楼市新闻', # news_house
'环球不含中国类别新闻', # news_world
'书籍文化历史类别新闻', # news_culture
'故事类别新闻', # news_story
'股票市场类别新闻', # news_stock
]
_tnews_options_list_str = "\n".join(f'{chr(ord("A") + i)}. {tnews_labels[i]}'
_tnews_options_list_str = '\n'.join(f'{chr(ord("A") + i)}. {tnews_labels[i]}'
for i in range(len(tnews_labels)))
_tnews_options_range_str = "，".join(f'“{chr(ord("A") + i)}”'
_tnews_options_range_str = '，'.join(f'“{chr(ord("A") + i)}”'
for i in range(len(tnews_labels)))
tnews_infer_cfg = dict(
......@@ -37,9 +37,9 @@ tnews_infer_cfg = dict(
type=PromptTemplate,
template=dict(round=[
dict(
role="HUMAN",
role='HUMAN',
prompt=
f"{{sentence}}\n请判断上述内容属于什么新闻？\n{_tnews_options_list_str}\n请从{_tnews_options_range_str}中进行选择。\n答：",
f'{{sentence}}\n请判断上述内容属于什么新闻？\n{_tnews_options_list_str}\n请从{_tnews_options_range_str}中进行选择。\n答：',
),
]),
),
......@@ -49,23 +49,23 @@ tnews_infer_cfg = dict(
tnews_eval_cfg = dict(
evaluator=dict(type=AccEvaluator),
pred_role="BOT",
pred_role='BOT',
pred_postprocessor=dict(type=first_capital_postprocess),
)
tnews_datasets = [
dict(
abbr="tnews-dev",
abbr='tnews-dev',
type=TNewsDataset_V2,
path="./data/FewCLUE/tnews/dev_few_all.json",
path='./data/FewCLUE/tnews/dev_few_all.json',
reader_cfg=tnews_reader_cfg,
infer_cfg=tnews_infer_cfg,
eval_cfg=tnews_eval_cfg,
),
dict(
abbr="tnews-test",
abbr='tnews-test',
type=TNewsDataset_V2,
path="./data/FewCLUE/tnews/test_public.json",
path='./data/FewCLUE/tnews/test_public.json',
reader_cfg=tnews_reader_cfg,
infer_cfg=tnews_infer_cfg,
eval_cfg=tnews_eval_cfg,
......
configs/datasets/FinanceIQ/FinanceIQ_gen_e0e6b5.py
View file @ aa2dd2b5
......@@ -40,16 +40,16 @@ for _name in financeIQ_all_sets:
ice_template=dict(
type=PromptTemplate,
template=dict(
begin="</E>",
begin='</E>',
round=[
dict(
role="HUMAN",
role='HUMAN',
prompt=
f"以下是关于{_ch_name}的单项选择题，请直接给出正确答案的选项。\n题目：{{question}}\nA. {{A}}\nB. {{B}}\nC. {{C}}\nD. {{D}}"
f'以下是关于{_ch_name}的单项选择题，请直接给出正确答案的选项。\n题目：{{question}}\nA. {{A}}\nB. {{B}}\nC. {{C}}\nD. {{D}}'
),
dict(role="BOT", prompt='答案是: {answer}'),
dict(role='BOT', prompt='答案是: {answer}'),
]),
ice_token="</E>",
ice_token='</E>',
),
retriever=dict(type=FixKRetriever, fix_id_list=[0, 1, 2, 3, 4]),
inferencer=dict(type=GenInferencer),
......@@ -62,13 +62,13 @@ for _name in financeIQ_all_sets:
financeIQ_datasets.append(
dict(
type=FinanceIQDataset,
path="./data/FinanceIQ/",
path='./data/FinanceIQ/',
name=_name,
abbr=f"FinanceIQ-{_name}",
abbr=f'FinanceIQ-{_name}',
reader_cfg=dict(
input_columns=["question", "A", "B", "C", "D"],
output_column="answer",
train_split="dev",
input_columns=['question', 'A', 'B', 'C', 'D'],
output_column='answer',
train_split='dev',
test_split='test'),
infer_cfg=financeIQ_infer_cfg,
eval_cfg=financeIQ_eval_cfg,
......
configs/datasets/FinanceIQ/FinanceIQ_ppl_42b9bd.py
View file @ aa2dd2b5
......@@ -40,17 +40,17 @@ for _name in financeIQ_all_sets:
type=PromptTemplate,
template={
answer: dict(
begin="</E>",
begin='</E>',
round=[
dict(
role="HUMAN",
prompt=f"以下是关于{_ch_name}的单项选择题，请直接给出正确答案的选项。\n题目：{{question}}\nA. {{A}}\nB. {{B}}\nC. {{C}}\nD. {{D}}"
role='HUMAN',
prompt=f'以下是关于{_ch_name}的单项选择题，请直接给出正确答案的选项。\n题目：{{question}}\nA. {{A}}\nB. {{B}}\nC. {{C}}\nD. {{D}}'
),
dict(role="BOT", prompt=f'答案是: {answer}'),
dict(role='BOT', prompt=f'答案是: {answer}'),
])
for answer in ["A", "B", "C", "D"]
for answer in ['A', 'B', 'C', 'D']
},
ice_token="</E>",
ice_token='</E>',
),
retriever=dict(type=FixKRetriever, fix_id_list=[0, 1, 2, 3, 4]),
inferencer=dict(type=PPLInferencer),
......@@ -61,13 +61,13 @@ for _name in financeIQ_all_sets:
financeIQ_datasets.append(
dict(
type=FinanceIQDataset,
path="./data/FinanceIQ/",
path='./data/FinanceIQ/',
name=_name,
abbr=f"FinanceIQ-{_name}",
abbr=f'FinanceIQ-{_name}',
reader_cfg=dict(
input_columns=["question", "A", "B", "C", "D"],
output_column="answer",
train_split="dev",
input_columns=['question', 'A', 'B', 'C', 'D'],
output_column='answer',
train_split='dev',
test_split='test'),
infer_cfg=financeIQ_infer_cfg,
eval_cfg=financeIQ_eval_cfg,
......
configs/datasets/GLUE_CoLA/GLUE_CoLA_ppl_77d0df.py
View file @ aa2dd2b5
......@@ -5,20 +5,20 @@ from opencompass.openicl.icl_evaluator import AccEvaluator
from opencompass.datasets import HFDataset
_hint = "The following are text classification questions. \n" \
"Please determine whether the following sentence is linguistically acceptable: " \
"0 means unacceptable, 1 means acceptable.\n"
_hint = 'The following are text classification questions. \n' \
'Please determine whether the following sentence is linguistically acceptable: ' \
'0 means unacceptable, 1 means acceptable.\n'
CoLA_infer_cfg = dict(
ice_template=dict(
type=PromptTemplate,
template="Sentence: {sentence}\nResult: {label}",
template='Sentence: {sentence}\nResult: {label}',
),
prompt_template=dict(
type=PromptTemplate,
template={
answer:
f"{_hint}</E>Sentence: {{sentence}}\nResult: {answer}"
f'{_hint}</E>Sentence: {{sentence}}\nResult: {answer}'
for answer in [0, 1]
},
ice_token='</E>',
......@@ -29,7 +29,7 @@ CoLA_infer_cfg = dict(
CoLA_eval_cfg = dict(evaluator=dict(type=AccEvaluator), )
CoLA_datasets = []
for _split in ["validation"]:
for _split in ['validation']:
CoLA_reader_cfg = dict(
input_columns=['sentence'],
......
configs/datasets/GLUE_MRPC/GLUE_MRPC_ppl_96564c.py
View file @ aa2dd2b5
......@@ -5,19 +5,19 @@ from opencompass.openicl.icl_evaluator import AccEvaluator
from opencompass.datasets import HFDataset
_hint = "The following are semantic matching questions. \n" \
"Please determine whether the following two sentences are semantically equivalent: " \
"0 means not equivalent, 1 means equivalent.\n"
_hint = 'The following are semantic matching questions. \n' \
'Please determine whether the following two sentences are semantically equivalent: ' \
'0 means not equivalent, 1 means equivalent.\n'
MRPC_infer_cfg = dict(
ice_template=dict(
type=PromptTemplate,
template="Sentence one: {sentence1}\nSentence two: {sentence2}\nResult: {label}",
template='Sentence one: {sentence1}\nSentence two: {sentence2}\nResult: {label}',
),
prompt_template=dict(
type=PromptTemplate,
template={
answer:
f"{_hint}</E>Sentence one: {{sentence1}}\nSentence two: {{sentence2}}\nResult: {answer}"
f'{_hint}</E>Sentence one: {{sentence1}}\nSentence two: {{sentence2}}\nResult: {answer}'
for answer in [0, 1]
},
ice_token='</E>',
......@@ -29,12 +29,12 @@ MRPC_eval_cfg = dict(evaluator=dict(type=AccEvaluator), )
MRPC_datasets = []
for _split in ["validation", "test"]:
for _split in ['validation', 'test']:
MRPC_reader_cfg = dict(
input_columns=['sentence1', 'sentence2'],
output_column='label',
train_split="train",
train_split='train',
test_split=_split
)
......
configs/datasets/GLUE_QQP/GLUE_QQP_ppl_250d00.py
View file @ aa2dd2b5
......@@ -5,19 +5,19 @@ from opencompass.openicl.icl_evaluator import AccEvaluator
from opencompass.datasets import HFDataset
_hint = "The following are semantic matching questions. \n" \
"Please determine whether the following two sentences are semantically duplicate: " \
"0 means not duplicate, 1 means duplicate.\n"
_hint = 'The following are semantic matching questions. \n' \
'Please determine whether the following two sentences are semantically duplicate: ' \
'0 means not duplicate, 1 means duplicate.\n'
QQP_infer_cfg = dict(
ice_template=dict(
type=PromptTemplate,
template="Sentence one: {question1}\nSentence two: {question2}\nResult: {label}",
template='Sentence one: {question1}\nSentence two: {question2}\nResult: {label}',
),
prompt_template=dict(
type=PromptTemplate,
template={
answer:
f"{_hint}</E>Sentence one: {{question1}}\nSentence two: {{question2}}\nResult: {answer}"
f'{_hint}</E>Sentence one: {{question1}}\nSentence two: {{question2}}\nResult: {answer}'
for answer in [0, 1]
},
ice_token='</E>',
......@@ -29,12 +29,12 @@ QQP_eval_cfg = dict(evaluator=dict(type=AccEvaluator), )
QQP_datasets = []
for _split in ["validation", "test"]:
for _split in ['validation', 'test']:
QQP_reader_cfg = dict(
input_columns=['question1', 'question2'],
output_column='label',
train_split="train",
train_split='train',
test_split=_split
)
......
configs/datasets/GaokaoBench/GaokaoBench_no_subjective_gen_4c31db.py
View file @ aa2dd2b5
......@@ -10,33 +10,33 @@ with read_base():
GaokaoBench_datasets = []
for folder, prompts in [
("Multiple-choice_Questions", MCQ_prompts),
("Fill-in-the-blank_Questions", FBQ_prompts),
('Multiple-choice_Questions', MCQ_prompts),
('Fill-in-the-blank_Questions', FBQ_prompts),
]:
for p in prompts:
reader_cfg = {
"input_columns": ["question"],
"output_column": "answer",
'input_columns': ['question'],
'output_column': 'answer',
}
infer_cfg = {
"ice_template": {
"type": PromptTemplate,
"template": {"round": [{"role": "HUMAN", "prompt": p["prefix_prompt"] + "{question}"}]},
"ice_token": "</E>",
'ice_template': {
'type': PromptTemplate,
'template': {'round': [{'role': 'HUMAN', 'prompt': p['prefix_prompt'] + '{question}'}]},
'ice_token': '</E>',
},
"retriever": {"type": ZeroRetriever},
"inferencer": {"type": GenInferencer, "max_out_len": 1024},
'retriever': {'type': ZeroRetriever},
'inferencer': {'type': GenInferencer, 'max_out_len': 1024},
}
eval_cfg = {
"evaluator": {"type": "GaokaoBenchEvaluator" + "_" + p["type"]},
"pred_role": "BOT",
'evaluator': {'type': 'GaokaoBenchEvaluator' + '_' + p['type']},
'pred_role': 'BOT',
}
dataset = {
"type": GaokaoBenchDataset,
"abbr": "GaokaoBench_" + p["keyword"],
"path": os.path.join("data", "GAOKAO-BENCH", "data", folder, p["keyword"] + ".json"),
"reader_cfg": reader_cfg,
"infer_cfg": infer_cfg,
"eval_cfg": eval_cfg,
'type': GaokaoBenchDataset,
'abbr': 'GaokaoBench_' + p['keyword'],
'path': os.path.join('data', 'GAOKAO-BENCH', 'data', folder, p['keyword'] + '.json'),
'reader_cfg': reader_cfg,
'infer_cfg': infer_cfg,
'eval_cfg': eval_cfg,
}
GaokaoBench_datasets.append(dataset)
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment