text stringlengths 0 2.4k |
|---|
ke a leboga, motheeletši, thobela. |
Tabakgolo iring ye e re |
Kgoro ya tlhabollo ya ekonomi tikologo le tša boeti ka mono profenseng, sephodisa sa naga le Limpopo gambling board legono di sepetša lenaneo la go fediša metšhene yeo e šomišwago go gambler ka fao go sego molaong. |
Ba tšea karolo ba ba kgobokeditše le go tšea metšhene ye lekgolo masome seswai tharo go tšwa di karolong tša go fapana tša profense ya Limpopo. |
Se ke maitekelo a department le batšeakarolo ago lwantšhana le di tiragalo tša go gambler ka fao go se go molaong. |
Selete sa Capricorn le sa Waterberg ke tšeo di itemogelwago palo ya godimo ya di tiragalo tša go gambler ka fao go sego molaong mme metšhene ye mentši ya go gambler e hweditšwe mabenkeleng e se ya ngwadišwa ebile e sena mangwalo. |
Mmoleledi Zayde Kalla o a hlaloša. |
company ya ESKOM e tsebagaditše gore e ka gapeletšega go tsentša tirišong magato a kgaotšo ya mohlagase ka bjako, se ke morago ga gore setiši sa go fehla mohlagase sa Kriel power se itemogele mathata a mohlagase mesong ya legono. |
ESKOM e re di unit tša go fehla mohlagase go ditiši tše dingwe tša go fehla mohlagase di ka boela sekeng mathapameng a legono efela go ka letelwa mathata a mangwe go akaretšwa le go kgaotša ga mohlagase. |
Moahlodi Mandisa Maya o kgethilwe bjalo ka mohlatša moahlodi mogolo o moswa ka mono nageng. |
MoPresitente Cyril Ramaphosa o tsebagaditše seo legono, mme o re Maya o na le mengwaga ye masomepedi bjalo ka ramolao. |
Ke mosadi wa mathomo wa go swara maemo a bo tlatša MoPresitente le a bo MoPresitente wa kgorotsheko ya godimo ya boipobolo. |
O tlile go thoma mošomo wa gagwe bjalo ka mohlatša moahlodi ka la pele kgweding ya Setemere. |
Tona ka kgorong ya di namelwa Fikile Mbalula o re ba sa tšwela Pele go amogela di kgopelo tša mašeleng a kimollo a baotledi ba dithekisi . |
Mbalula o re boradithekisi ba eletšwa go ingwadiša gore ba tle ba hwetše mašeleng a bona a kimollo. |
Mbalula o re industry ya dithekisi Ke enngwe ya di industry tšeo di amegilego kudu ge Naga e be e le ka gare Ga seemo sa masetlapelo. |
o re boradithekisi ba filwe go fihla ka la masometharo tee kgweding ya hlakola ngwageng woo o tlago go dira di kgopelo tša mašeleng ao. |
Mbalula o a hlaloša. |
Ke ruma di taba tše mme ke dira bjalo ka go tsopola tabakgolo mme yona e re |
company ya ESKOM e tsebagaditše gore e ka gapeletšega go tsentša tirišong magato a kgaotšo ya mohlagase ka bjako. |
ke a leboga, motheeletši wa radio turf, thobela. |
Tabakgolo iring ye e re |
Moahlodi Mandisa Maya o kgethilwe bjalo ka mohlatša moahlodi mogolo o moswa ka mono nageng. |
MoPresitente Cyril Ramaphosa o tsebagaditše seo legono, mme o re Maya o na le mengwaga ye masomepedi bjalo ka ramolao. |
Ke mosadi wa mathomo wa go swara maemo a bo tlatša MoPresitente le a bo MoPresitente wa kgorotsheko ya godimo ya boipobolo. |
O tlile go thoma mošomo wa gagwe bjalo ka mohlatša moahlodi ka la pele kgweding ya Setemere. |
Tona ka kgorong ya di namelwa Fikile Mbalula o re ba sa tšwela Pele go amogela di kgopelo tša mašeleng a kimollo a baotledi ba dithekisi . |
Mbalula o re boradithekisi ba eletšwa go ingwadiša gore ba tle ba hwetše mašeleng a bona a kimollo. |
Mbalula o re industry ya dithekisi Ke enngwe ya di industry tšeo di amegilego kudu ge Naga e be e le ka gare Ga seemo sa masetlapelo. |
o re boradithekisi ba filwe go fihla ka la masometharo tee kgweding ya hlakola ngwageng woo o tlago go dira di kgopelo tša mašeleng ao. |
Mbalula o a hlaloša. |
MoPresitente wa mokgahlo wa African National Congress Cyril Ramaphosa o iphetoletše go di polelo tšeo di tšweleditšego Ke MoPresitente wa maloba wa Naga Thabo Mbeki, malebana le mathata ao Naga e lebanego le o na. |
Se Ke morago ga gore bekeng ya go feta, tirelong ya segopotšo sa bophelo bja mohlatša mongwaledipharephare wa African National Congress Jessie Duarte, Mbeki a re boetapele bja Naga Bo thekgile ditho, Ga ba tle ka leano la go thuša go ditlhotlo tšeo Naga e lebanego le tšona. |
Mbeki o re o tšhošwa Ke gore bomenetša Bo šitiša Naga go fihlela tlhabologo ka mono nageng. |
Ramaphosa o re go swanetšwe go gopole gore ditlhotlo tše a se tšeo tša go thoma maabane, Ke tšeo di thomilego mengwageng e mentši ya go feta. |
Ramaphosa o a hlaloša. |
Bašomi ba lekala la motšhelo ka mono nageng la SARS Bao ba ingwadišitšego le mokgahlo wa Public servants Association (PSA), ba sa tšwela Pele ka mogwanto wa bona. |
Bašomi bao ba re ba tlo tšwela Pele ka mogwanto woo goba go fihlela ge MoPresitente wa Naga Cyril Ramaphosa a sekegela dillo tša bona tsebe. |
Se Ke morago ga gore bašomi bao e ka bago ba sekete, bao ba ingwadišitšego le PSA ba tsenele magato a boipelaetšo pejana kgweding ye ba ipelaetša ka megolo ya bona le tše dingwe . |
Bare ba file Ramaphosa go fihla ka la pele kgweding yeo e tlago, Mme ge aka se ba fetole go tla nkga go se gwa bola. |
Ke ruma ka go tsopola tabakgolo yona e re. |
Moahlodi Mandiša Maya o kgethilwe bjalo ka mohlatša moahlodi mogolo o moswa ka mono nageng. |
MoPresitente Cyril Ramaphosa o tsebagaditše seo legono, mme o re Maya o na le mengwaga ye masomepedi bjalo ka ramolao. |
Ke mafalelo a di taba tša iri ya bo lesomepedi, ke boya gape ka iri ya pele. |
ke a leboga, tabakgolo iring ye |
Moeta pele wa Democratic Alliance ka Bela Bela Desiree Van der Walt o bontšhitše gore meyara wa masepala Gloria Seleka o ganne go amogela lengwalo la bona la di ngongorego. |
Se ke morago ga gore Democratic Alliance le batšeakarolo ba go fapana ba tsenele mogwanto go leba di kantorong tša masepala e le ge ba lla ka kabo ya ditirelo go e tša hlokego ya meetse le kgaotšo ya mohlagase magareng ga tše dingwe . |
Democratic Alliance e re e tlo tšwela Pele go lwa ntwa ya letsogo kgaoga go fihlela ge dillo tša bona di sekegelwa tsebe. |
Van der Walt o a hlaloša. |
Tona wa maphodisa Bheki Cele o lokolotše di palopalo tša bosenyi Palamenteng legono. |
Cele o bontšhitše gore melato ya di tlaišo tša bong e hlatlogile ka di phesente tše lesome tharo fegelwana šupa ge di bapetšwa le tša ngwageng wa go feta. |
O bontšhitše gape gore di palopalo go šeditšwe di polao e hlatlogile ka di phesente tše masome pedi pedi mo di kgweding tše tharo. |
Ge a tšwela pele o laeditše gore di taba tše botse ke gore ge e le melato go šeditšwe go thuba ka malapeng e theogetše tlasana, mola di phesente tša melato ya maitekelo a polao e le masome tlhano. |
Cele o be a emwe nokeng ke major general Norman Sekhukhune mo a bontšhitšego gore ba tlo leka go šogana le di phesente tšeo di hlatlogilego. |
Sekhukhune o a hlaloša. |
Ba mošate wa Bjatladi, Segopje ga Mamabolo ba ipiletša go badudi gore ba hlokomele digongwana tša batho bao ba tsenago motse ka motse ba kgopela mašeleng ka leina la mošate. |
Johannes Mamabolo e lego mokgomana wa mošate o re ba lemoša gore taba yeo ga e gona, se ke morago ga gore ba bangwe ba badudi ba lle ka bothata bjoo bja go jabetša batho mašeleng. |
Mamabolo o tšwetše pele ka gore bao ba itemogetšego bothata bjo ba bule molato maphodi seng. |
Mamabolo o a hlaloša. |
Lefapha la dithuto tša maphelo ka mono yunibesithing ya Limpopo le gare ka go sepetša lenaneo Leo le nepilego go hlompha bahlankedi ba tša maphelo le go ba hlohleletša gore ba etiše pele lephelo la bona la go Amana le monagano ke gore mental health. |
Lenaneo le le nepile gape go lemoša badudi gore dingaka le tšona di fela di lebagana le ditlhotlo go etša motho mang le mang. |
Lenaneo le tšea karolo ka mokgwa woo go wona baithuti le bahlankedi lefapheng Leo ba tsenetšego mosepelo go tloga ka mono yunibesithing go ya bookelong bja Mankweng. |
Dr tšakani Sumbani go tšwa lefapheng la dithuto tša bongaka ka mono yunibesithing ebago morulaganyi wa mananeo o re bahlankedi ba ba tša maphelo ba kgatha karolo ye kgolo malebana le go hlokomela balwetši mme o rile go bohlokwa gore le bona ba hlokomele lephelo la bona. |
Mmoleledi wa ditirelo tša balaolasephethephethe ka tlase ga masepala wa Polokwane Kgaogelo Mphago o re di senthara tša go aba ditirelo tša mangwalo ago otlela ebago ya Mankweng, Polokwane le Aganang di tlaba di šoma moo mafelelong a beke. |
Mphago o re di senthara tšeo di ya go bulwa ka iri ya bo seswai go fihla ka iri ya bone. |
O re ba ipeletša go baotledi bao ba nyakago go tsošološa mangwalo a bona ago otlela le go dira metšhelo go ikgaša fao. |
Mphago o a hlaloša. |
Ke a leboga, motheeletši thobela. |
Tabakgolo iring ye e re Gona le di kgonagalo tša gore moeta pele wa union ya NUMSA Andrew Chirwa le mongwaledi pharephare Irvin Jim ba ka golegwa morago ga gore ba tshele molao wa kgorotsheko. |
Se ke morago ga gore maloko a NUMSA a tšwele pele ka congress ya bo lesome tee ya go kgetha boetapele bekeng ya go feta le ge kgorotsheko e be e ileditše seo. |
Ditokomane tšeo tša kgorotsheko di rometšwe morago ga gore mohlatša moeta pele wa bobedi wa maloba ka go union yeo Ruth Ntlokose a leke go hlotla sephetho sa union sa go mo fega mmogo le maloko a mangwe. |
Le ge go le bjalo union yeo e ganane le molato woo e pharwago ka wona ka ge e re e latetše taelo ya kgorotsheko. |
Lekala la mašoka ka mono nageng la SANPARKS le lokolotše matšatši ao le tlogo sepetša lenaneo la ngwaga ka ngwaga leo baeti ba fiwago go tsena mašokeng a di phoofolo ntle le tefo. |
SANPARKS e re lenyaga lenaneo leo le ya go sepetšwa go thoma ka la 11 go fihla ka la 16 kgweding yeo e tlago mašokeng a lona a 22 ka mono nageng. |
Lenaneo le sa le le thakgolwa ngwageng wa 2006. |
Mohlankedi wa di kgokaganyo kago SANPARKS Rey Thakuli o a hlaloša. |
Baotledi ba difatanaga ba tlo imologa go se nene ka ge go letetšwe gore theko ya petrol e theogele tlase ka ranta le di sente tše masometharo pedi litara bošegogare. |
Theko ya diesel e ya go theoga ka di sente tše masomesenyane tee litara mme paraffin e ya go theogela tlase ka ranta le di sente tše masome nne nne. |
Tirelo ya poloko ya meyara wa masepala wa Collins Chabane Moses Maluleke e gare e ya kgatlampana le patlelong la di papadi la Saselamani ka ntle Ga Malamulele. |
Maluleke o thuntšhitšwe le go bolawa di bekeng tše pedi tša go feta motšaneng wa Sikundu ke bagononelwa ba bararo bao ba bego ba itlhamile ka di thunya. |
O be a na le morwa wa gagwe wa mengwaga ye 18 yo a itemogetšego di kgobalo yena o lokolotšwe bookelong yesterday. |
Tonakgolo Stanley Chupu Mathabatha le Makgotlaphethišo a mangwe go sa tlogelwe maloko a mokgahlo wa African National Congress a bothane fao go laelana le meyara Maluleke woo gantši a bego a bitšwa Big Moss. |
Tonakgolo Mathabatha o re maphodisa a dira mošomo o mo botse ka mono profenseng malebana le go golega bagononelwa eupša di kgorotsheko di fa balatofatšwa kotlo ye bonolo. |
Ke a leboga, motheeletši thobela |
tabakgolo iring ye e re |
Mongwaledipharephare wa union ya go emela bašomi ba go šoma ka tšhipi ya NUMSA Irvin Jim le MoPresitente wa union yeo Andrew Chirwa ba kgethilwe gape Maemong a bona nakong ya Khonferense ya bosetšhaba ya union ka Cape town maabane. |
Khonferense Yeo Ke Yeo e tsenetšego Ke di kemedi tše sekete Mme go letetšwe gore e rungwe gosasa. |
NUMSA e sepetša khonferense yeo le ge kgorotsheko eba laetše gore ba se e sepetše. |
Phumzi Maqungo o kgethilwe bjalo ka Ramatlotlo pharephare Maemong a bosetšhaba. |
MoPresitente wa union Andrew Chirwa o re moahlodi kgorongtsheko ya mešomo Ga ana kwešišo Yeo e tletšego go molaotheo wa union NUMSA Ke ka fao a fihleletšego sephetho Seo a se tšerego Chirwa o a hlaloša. |
Lekala la motšhelo ka mono nageng la SARS le re le ikemišeditše go tiiša letsogo ge le koleka mašeleng a motšhelo Mo sehleng se se tlago, kudu go badudi ba Naga Bao ba Lego di nageng tša ka ntle. |
SARS e re e tlo fa šedi e kgolo go badudi Bao ba sego ba ingwadiša go SARS ge ba huduga ka mono nageng. |
Ramolao wa go šetšana le metšhelo ya badudi ka di nageng tša ka ntle Martin Bezuidenhout o re o eletša badudi gore ge ba hudugela ka di nageng tša ka ntle ba ikgokaganye le SARS go hwetša tshedimošo le thušo ya maleba. |
Bezuidenhout o a hlaloša. |
Le legono department ya tlhabollo ya Leago ka mono profenseng e sa sepetša mananeo a Yona ago keteka kgwedi ya MoPresitente wa maloba Nelson Mandela, ka go fa ditirelo tša thobamatshwalo go ba malapa a bo batšwasehlabelo ba di tlaišo tša bong mono tikologong ya Mankweng. |
Molekgotlaphethiši wa department ye ka mono profenseng Dr Namane Dickson Masemola o re legono ba tlabe ba sepetša mananeo a ka lekaleng la Andrew Mehlape ka di kantorong tša department ya tlhabollo ya Leago mono Mankweng go boledišana le Banna gore ba tlogele go gataka di tokelo tša basadi le Bana. |
Masemola o re ba tlo netefatša gore basadi le Bana ba sepela ba bolokegile Mo mebileng fao a rego Banna ba swanetše goba bona ba netefatšago polokego le tšhireletšo ya bona. |
Masemola o a hlaloša. |
Tona ka Kgorong ya diminerale le Energy Gwede Mantashe o re go tlo fokotšwa di theko tša makhura kgweding Yeo e tlago. |
Mantashe o re go tlo lokolwa sengwalwa sa maleba Pele Ga ge beke ye e fela go tsebagatša ka seo. |
End of preview. Expand in Data Studio
language: - nso - en license: cc-by-nc-4.0 pretty_name: "SepNews: Sepedi News Corpus" task_categories: - text-generation tags: - sepedi - african-languages - low-resource-languages - natural-language-processing - computational-linguistics - news-corpus - language-identification - code-mixing - sepedi-english
SepNews: Sepedi News Corpus
Dataset Description
Overview
SepNews is a collection of Sepedi news texts compiled from news bulletin scripts from a local community radio station in Limpopo, South Africa. The corpus was developed to support research and development in natural language processing (NLP), particularly for Sepedi, a low-resource African language. The corpus consists of two collections, SepNews-1 and SepNews-2, collected during two different periods:
- SepNews-1: June 2022 to April 2023.
- SepNews-2: May 2023 to November 2023.
The two collections provide news texts from different periods and support research on Sepedi language processing, language modelling, lexical analysis, and Sepedi-English code-mixing.
SepNews-1 is intended for training and development, whereas SepNews-2 is reserved as a held-out test set. This separation supports evaluation on data collected during a later period than the training and development data.
Dataset Summary
| Property | Description |
|---|---|
| Dataset name | SepNews |
| Languages | Sepedi and English |
| Language codes | nso, en |
| Domain | News |
| Data type | Text |
| Number of collections | 2 |
| Collection periods | June 2022 – November 2023 |
| Primary application | Low-resource NLP |
| Specific focus | Sepedi-English code-mixing |
| Licence | CC BY-NC 4.0 |
Dataset Statistics
The following table presents the main statistics for the two collections.
| Statistic | SepNews-1 | SepNews-2 |
|---|---|---|
| Collection period | June 2022 – April 2023 | May – November 2023 |
| Number of sentences | 15,583 | 3,520 |
| Number of tokens | 418,651 | 101,758 |
| Number of unique tokens | 16,143 | 6,420 |
| English token proportion | 3.85% | 11.92% |
The English token proportion indicates the presence of English words in the Sepedi news texts and reflects the occurrence of Sepedi-English code-mixing.
Dataset Structure
The dataset is organised into two collections:
| Collection | Intended use | Description |
|---|---|---|
| SepNews-1 | Training and development | News texts collected between June 2022 and April 2023. |
| SepNews-2 | Held-out testing | News texts collected between May and November 2023. |
SepNews-2 is reserved exclusively for testing and independent evaluation.
To preserve the integrity of the test set, researchers should not use SepNews-2 for:
- Model training or continued pretraining.
- Fine-tuning or instruction tuning.
- Training or validation of language identification systems.
- Vocabulary adaptation or model development.
- Hyperparameter optimisation or other development activities.
SepNews-2 should only be used for final evaluation after model development has been completed using appropriate training and development data.
This temporal separation is intended to provide a more realistic assessment of model performance on news text from a later collection period.
Data Collection
The datasets developed in this study were sourced from a local community radio station. To avoid the time-consuming process of manual transcription, preprepared news bulletin scripts were obtained directly from the editorial team. These bulletins were produced under editorial supervision and read by Sepedi news presenters during the broadcast, ensuring consistency in language use and style.
Intended Uses
SepNews is intended primarily for academic and non-commercial research in natural language processing and computational linguistics.
Potential research applications include:
- Language modelling: Pretraining, continued pretraining and evaluation of Sepedi language models.
- Low-resource NLP: Development and evaluation of NLP methods for under-resourced African languages.
- Code-mixing analysis: Investigation of Sepedi-English code-mixing patterns in news text.
- Language identification: Development and evaluation of word-level and sentence-level language identification methods.
- Lexical analysis: Investigation of vocabulary distribution and lexical characteristics in Sepedi news.
- Text processing: Development and evaluation of text normalisation and other language-processing techniques.
Researchers using the corpus for language modelling are encouraged to maintain the designated training/development and test separation.
Limitations and Considerations
The following limitations should be considered when using this dataset:
- Domain coverage: The corpus consists of news bulletin scripts and does not represent all forms of written or spoken Sepedi.
- Language variation: The dataset may not capture the full range of Sepedi dialectal, orthographic and stylistic variation.
- Code-mixing: The proportion and distribution of English words reflect the content and language practices represented in the collected news scripts.
- Temporal coverage: The corpus covers two specific collection periods between June 2022 and November 2023.
- Generalisation: Models trained on this dataset may not generalise equally well to other domains, registers or forms of Sepedi language use.
The dataset should therefore be interpreted in the context of its news-domain composition and collection periods.
Licensing Information
This dataset is made available under the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) licence.
- Appropriate credit must be given to the dataset authors.
- A link to the CC BY-NC 4.0 licence must be provided.
- Any modifications made to the dataset must be indicated.
- The dataset may not be used for commercial purposes without additional permission from the relevant rights holders.
Citation
If you use SepNews in your research, please cite the dataset as follows:
Dataset citation:
A research paper describing the dataset and the Sepedi Language Validation Pipeline (SLVP) is currently under review. Full publication details and the corresponding citation will be provided following publication.
Related research publication:
Please acknowledge the dataset when reporting research results obtained using SepNews.
Acknowledgements
Dataset Maintainers and Authors
Contact
- Downloads last month
- 64