mopatik/PuoBERTa-offensive-detection-v1
Updated
Dataset Preview
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed because of a cast error
Error code: DatasetGenerationCastError
Exception: DatasetGenerationCastError
Message: An error occurred while generating the dataset
All the data files must have the same columns, but at some point there are 1 new columns ({'TARGET LABEL'}) and 1 missing columns ({'TARGET'}).
This happened while the csv dataset builder was generating data using
hf://datasets/mopatik/setswana-offensive-977/testing_data-tagged.csv (at revision fa59fb4adac346065df5279ca52370458e386325)
Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1831, in _prepare_split_single
writer.write_table(table)
File "/usr/local/lib/python3.12/site-packages/datasets/arrow_writer.py", line 714, in write_table
pa_table = table_cast(pa_table, self._schema)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2272, in table_cast
return cast_table_to_schema(table, schema)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2218, in cast_table_to_schema
raise CastError(
datasets.table.CastError: Couldn't cast
TEXT: string
TARGET LABEL: string
-- schema metadata --
pandas: '{"index_columns": [{"kind": "range", "name": null, "start": 0, "' + 493
to
{'TEXT': Value('string'), 'TARGET': Value('string')}
because column names don't match
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1339, in compute_config_parquet_and_info_response
parquet_operations = convert_to_parquet(builder)
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 972, in convert_to_parquet
builder.download_and_prepare(
File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 894, in download_and_prepare
self._download_and_prepare(
File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 970, in _download_and_prepare
self._prepare_split(split_generator, **prepare_split_kwargs)
File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1702, in _prepare_split
for job_id, done, content in self._prepare_split_single(
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1833, in _prepare_split_single
raise DatasetGenerationCastError.from_cast_error(
datasets.exceptions.DatasetGenerationCastError: An error occurred while generating the dataset
All the data files must have the same columns, but at some point there are 1 new columns ({'TARGET LABEL'}) and 1 missing columns ({'TARGET'}).
This happened while the csv dataset builder was generating data using
hf://datasets/mopatik/setswana-offensive-977/testing_data-tagged.csv (at revision fa59fb4adac346065df5279ca52370458e386325)
Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
TEXT string | TARGET string |
|---|---|
letlharapa le le kgobokanelwang ke dinonyane tsotlhe | Non-offensive |
banna ga ba na thwadi ba bonwe ke mmamariga | Non-offensive |
o maswe | Offensive |
ba re o kgaetse malome | Offensive |
buisa temana malebana le go fedisa botsalano | Non-offensive |
o ka se bone last born e didimetse di makgakga bogoma dingalo di rata dijo di rata ntwa di conclusion tseo | Offensive |
ga go na konteraka madi a tla go fitlhelela ka bonako mo akhaonteng ya gago | Offensive |
go nna modidi wa nta e e motopo | Non-offensive |
waitse basadi hle re di bari monna santse a lwelwa o tshosetsa yoo mongwe sehema ke wena | Offensive |
dipalopalo tsa akhaonto ya gago di bontsha madi a a sa tlwaelegang | Offensive |
dijo tse di monate di a re kgatlha | Non-offensive |
go tlhomola pelo | Non-offensive |
nxla ke tenegile | Offensive |
go itaya kgomo lonaka | Non-offensive |
ke tla go tlhobaetsa | Offensive |
kgwele ya dinao | Non-offensive |
soca e ikaelela go | Non-offensive |
ke tla go tlhorontsha boroko | Offensive |
o tlhoka tlhaloganyo ke eng o sa kgone go sekaseka dilo sentle | Offensive |
buruburu rradijase | Non-offensive |
re tlhokometse diphetogo tse di sa tlwaelegang mo akhaonteng ya gago re fe dinomoro tsa gago | Offensive |
ba tshwanetse go tlhabolola | Non-offensive |
ba ba buelang dikgageng magakabe a a ba bona | Non-offensive |
re bona ditiragalo tse di sa tlwaelegang mo akhaontong ya twitter. kopa o netefatse ka go tlatsa data ya gago | Offensive |
ha o sena madi a ticketemella gompieno yaana o ye go iphitha mo stadium o lese matepe | Offensive |
boaka le maaka tshele o di rata tota | Offensive |
dintlha tsa gago tsa payment di tlhoka go netefatswa. tobetsa fano go ntsha di-passwords le banking infoTRIGGER> | Offensive |
areyeng gaborone | Non-offensive |
hewe hewe ke na le monna kante o raya legatlapa lele tshabang go tswela ko ntle bosigo la gorota le lala le rotela o emereng | Offensive |
basadi ga ba tshwanela go tsaya ditshwetso mo kerekeng | Offensive |
waitse matagwa one ga ana order | Offensive |
guys i need your advice ke batla go ya ko jarateng e tona keye go dumedisa bo kgaitsadiake le nkuku ba baloi | Offensive |
diso | Offensive |
kgomola lodi o je kota | Non-offensive |
go se neele batshwarwa tiro go kotsi | Non-offensive |
kgobokanya tshedimosetso e e latelang | Non-offensive |
tsweetswee fa tshedimosetso ya omang, banka , le go tlatsa fomo go tsaya lotlhapelo ya madi. | Offensive |
makgakga a moruti nxla | Offensive |
bontlenyane bo seng bosi bo na le dibelebejane | Non-offensive |
nnyo ya gago e phala mosima wa thakadu | Offensive |
go logela motho segwe | Non-offensive |
basadi ba bolola ntwa ba sa e ya ditlholwa | Non-offensive |
re leboga ramasedi ka sengwe le sengwe | Non-offensive |
go nyala lefufa | Non-offensive |
join le rona mo bitcoin investment. phela go le gompieno o bona p20,000 mo dibekeng tse 2 | Offensive |
re a itse gore re tlhophile tsela e e thatanyana go ya kgolong le tlhabololo | Non-offensive |
sematla sa bofelo | Offensive |
ke tla go gasagasa | Offensive |
mabelete a kwa ke mathata | Offensive |
nna ga ke batle go bua thata a video eo e goroge ko whatsapp | Offensive |
lephata le tlaa buisana le go tsaya tshwetso ya gore le dire eng | Non-offensive |
ke santse ke dueletse batho go go batla | Offensive |
tlhokomeliso ya molao: o kolota lekgetho tlatsa tshedimosetso go lefa p1,200 mo irs ka letsatsi le le fa tlase. | Offensive |
ditekanyo di akaretsa gore | Non-offensive |
go nna morui wa kgomo e telele | Non-offensive |
mme re ka se tlhole re kgona go leta | Non-offensive |
batho ba ba polotiki ba ke magodu o seka wa ba tsaya tsia | Offensive |
tswa foo o ise o ntene sematla ke wena | Offensive |
akhaonte ya gago e tla tswalwa fa o sa arabe jaana | Offensive |
re tsere tshelete e e tswang mo akhaontong ya gago go lefa debit. netefatsa gore transaction e ne e le ya gago | Offensive |
tota mona a o seso | Offensive |
go anetswa dipodi | Non-offensive |
go ikepela lemena | Non-offensive |
o tsaya kae bogoma jwa go smiler o nale rusi mo menong | Offensive |
bontšhe ga ba tlhakanele khutla | Non-offensive |
ke tla go latlha | Offensive |
kgopolo ya me ga e a fetolwa go ya ka ntlha eno | Non-offensive |
tobetsa ha go tsaya p12000 o o winne mo patising ya botswana skills fund. | Offensive |
bogosi bo a taga | Non-offensive |
le nne le tsenya le diaka tsa lona re di bone | Offensive |
basimane le basetsana ba kgabile thata | Non-offensive |
ngwanyana yo ga a itse sepe ke mo direla fela ka dinako tsotlhe | Offensive |
tsamayang le morena | Non-offensive |
monna yo o humang bosigo motshegare a humanega | Non-offensive |
o tla duelela sotlhe se o se dirileng | Offensive |
nnaka gare wena o setlhodi | Offensive |
ka jalo re rata go go leboga thatathata | Non-offensive |
re setse re tsentse bwp 500 mo akhaonteng ya gago re fe nomoro ya netefatso | Offensive |
ga e khuetse puso fela e khuetsa le lephata la poraefete | Non-offensive |
go apeela thukgwinoko le mala | Non-offensive |
fa re filwe matlhakore a le 3 re ka rarabolola dikhutlo dingwe le dingwe tse 3 | Non-offensive |
a o ka nagana | Non-offensive |
re batla go fetisetsa p1.5 million go tswa kwa nigeria o tla fiwa p150,000 | Offensive |
batho ba tshwanetse go lemoga diphoso tsa bone | Non-offensive |
ke moemedi wa mascom money ke tlhoka nomoro ya gago ya sephiri | Offensive |
baloi ba ga lona ba feditse ka wena | Offensive |
kgale ke didimetse le bua matlakala ka nna | Offensive |
go tshwara tau ka mangana | Non-offensive |
monna wa phakela a le fa ditedu ke nama molomo ke lerapo | Non-offensive |
dirisa dikaedi tse di latelang mo dipuisanong tsa gago | Non-offensive |
bana ba tadi ba itsane ka mereto | Non-offensive |
rotlhe re itse setlhabelo se a se neetseng | Non-offensive |
mariga a tlhola serame se se thata | Non-offensive |
dirisa dichelete tea gago sentle | Non-offensive |
se se phimotswe ke goreng jalo | Non-offensive |
ngaka ya setso e tla go thusa | Non-offensive |
bua gore seno ke tekatekano e e ntle le e e seng ntle | Non-offensive |
bogosi ga bo tlolwe e se letsoku | Non-offensive |
botswerere jwa modisa ke go di tlhatlhela | Non-offensive |
ba ne ba sa itse ka malwetse a amanang le tiriso ya dikumo tsa motsoko | Non-offensive |
End of preview.
Setswana Offensive Language Dataset (977)
Tasks: Binary offensive language detection (0 = Non-offensive, 1 = Offensive)
Language: Setswana (tsn; Botswana/South Africa/Namibia)
Size: 977 instances (477 offensive; 500 non-offensive)
Splits
- Protocol: 80/20 holdout (tag-free, unaugmented). 5-fold CV on the 80% train.
- Evaluation: Holdout tag-free, mirrors deployment.
Annotation
- Labels: {0: Non-offensive, 1: Offensive}
- Spans: Semantic trigger phrases (train-time experiments).
- Inter-annotator agreement: κ = 0.86 (substantial).
Files
train.csv,test.csv(orfull.csvwith asplitcolumn)DATASTATEMENT.md,LICENSE, (optional)checksums.sha256
Load
from datasets import load_dataset
ds = load_dataset("mopatik/setswana-offensive-977")
print(ds)
# If using CSVs:
# ds = load_dataset("csv", data_files={"train": "train.csv", "test": "test.csv"})
- Downloads last month
- 14