mirror of
https://github.com/open-compass/opencompass.git
synced 2025-05-30 16:03:24 +08:00
d8354fe5d8
* [Dataset] add GLUE_CoLA dataset * [update] use HFDataset to load glue/cola dataset * update --------- Co-authored-by: gaotongxiao <gaotongxiao@gmail.com>
51 lines
1.5 KiB
Python
51 lines
1.5 KiB
Python
from opencompass.openicl.icl_prompt_template import PromptTemplate
|
|
from opencompass.openicl.icl_retriever import FixKRetriever
|
|
from opencompass.openicl.icl_inferencer import PPLInferencer
|
|
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"
|
|
|
|
CoLA_infer_cfg = dict(
|
|
ice_template=dict(
|
|
type=PromptTemplate,
|
|
template="Sentence: {sentence}\nResult: {label}",
|
|
),
|
|
prompt_template=dict(
|
|
type=PromptTemplate,
|
|
template={
|
|
answer:
|
|
f"{_hint}</E>Sentence: {{sentence}}\nResult: {answer}"
|
|
for answer in [0, 1]
|
|
},
|
|
ice_token='</E>',
|
|
),
|
|
retriever=dict(type=FixKRetriever),
|
|
inferencer=dict(type=PPLInferencer, fix_id_list=[17, 18, 19, 20, 21]))
|
|
|
|
CoLA_eval_cfg = dict(evaluator=dict(type=AccEvaluator), )
|
|
|
|
CoLA_datasets = []
|
|
for _split in ["validation"]:
|
|
|
|
CoLA_reader_cfg = dict(
|
|
input_columns=['sentence'],
|
|
output_column='label',
|
|
test_split=_split
|
|
)
|
|
|
|
CoLA_datasets.append(
|
|
dict(
|
|
abbr=f'CoLA-{_split}',
|
|
type=HFDataset,
|
|
path='glue',
|
|
name='cola',
|
|
reader_cfg=CoLA_reader_cfg,
|
|
infer_cfg=CoLA_infer_cfg,
|
|
eval_cfg=CoLA_eval_cfg
|
|
)
|
|
)
|