{"id":4265,"date":"2025-09-14T01:50:39","date_gmt":"2025-09-14T01:50:39","guid":{"rendered":"https:\/\/salarydistribution.com\/machine-learning\/2025\/09\/14\/reaching-across-the-isles-uk-llm-brings-ai-to-uk-languages-with-nvidia-nemotron\/"},"modified":"2025-09-14T01:50:39","modified_gmt":"2025-09-14T01:50:39","slug":"reaching-across-the-isles-uk-llm-brings-ai-to-uk-languages-with-nvidia-nemotron","status":"publish","type":"post","link":"https:\/\/salarydistribution.com\/machine-learning\/2025\/09\/14\/reaching-across-the-isles-uk-llm-brings-ai-to-uk-languages-with-nvidia-nemotron\/","title":{"rendered":"Reaching Across the Isles: UK-LLM Brings AI to UK Languages With NVIDIA Nemotron"},"content":{"rendered":"
\n\t\t<\/span> <\/span> <\/span><\/span><\/p>\n

Celtic languages \u2014 including Cornish, Irish, Scottish Gaelic and Welsh \u2014 are the U.K.\u2019s oldest living languages. To empower their speakers, the UK-LLM sovereign AI<\/a> initiative is building an AI model based on NVIDIA Nemotron<\/a> that can reason in both English and Welsh, a language spoken by about 850,000 people<\/a> in Wales today.<\/p>\n

Enabling high-quality AI reasoning in Welsh will support the delivery of public services including healthcare, education and legal resources in the language.<\/p>\n

\u201cI want every corner of the U.K. to be able to harness the benefits of artificial intelligence. By enabling AI to reason in Welsh, we\u2019re making sure that public services \u2014 from healthcare to education \u2014 are accessible to everyone, in the language they live by,\u201d said U.K. Prime Minister Keir Starmer. \u201cThis is a powerful example of how the latest AI technology, trained on the U.K.\u2019s most advanced AI supercomputer in Bristol, can serve the public good, protect cultural heritage and unlock opportunity across the country.\u201d<\/p>\n

The UK-LLM project, established in 2023 as BritLLM and led by University College London, has previously released two models for U.K. languages. Its new model for Welsh, developed in collaboration with Wales\u2019 Bangor University and NVIDIA, aligns with Welsh government efforts to boost the active use of the language, with the goal of achieving a million speakers by 2050 \u2014 an initiative known as Cymraeg 2050<\/a>.<\/p>\n

U.K.-based AI cloud provider Nscale will make the new model available to developers through its application programming interface. \"\"<\/p>\n

\u201cThe aim is to ensure that Welsh remains a living, breathing language that continues to develop with the times,\u201d said Gruffudd Prys, senior terminologist and head of the Language Technologies Unit at Canolfan Bedwyr, the university\u2019s center for Welsh language services, research and technology. \u201cAI shows enormous potential to help with second-language acquisition of Welsh as well as for enabling native speakers to improve their language skills.\u201d<\/p>\n

This new model could also boost the accessibility of Welsh resources by enabling public institutions and businesses operating in Wales to translate content or provide bilingual chatbot services. This can help groups including healthcare providers, educators, broadcasters, retailers and restaurant owners ensure their written content is as readily available in Welsh as they are in English.<\/p>\n

Beyond Welsh, the UK-LLM team aims to apply the same methodology used for its new model to develop AI models for other languages spoken across the U.K. such as Cornish, Irish, Scots and Scottish Gaelic \u2014 as well as work with international collaborators to build models for languages from Africa and Southeast Asia.<\/p>\n

\u201cThis collaboration with NVIDIA and Bangor University enabled us to create new training data and train a new model in record time, accelerating our goal to build the best-ever language model for Welsh,\u201d said Pontus Stenetorp, professor of natural language processing and deputy director for the Centre of Artificial Intelligence at University College London. \u201cOur aim is to take the insights gained from the Welsh model and apply them to other minority languages, in the U.K. and across the globe.\u201d<\/p>\n

Tapping Sovereign AI Infrastructure for Model Development\u00a0<\/b><\/h2>\n

The new model for Welsh is based on NVIDIA Nemotron<\/a>, a family of open-source models that features open weights, datasets and recipes. The UK-LLM development team has tapped the 49-billion-parameter Llama Nemotron Super model and 9-billion-parameter Nemotron Nano model, post-training<\/a> them on Welsh-language data.<\/p>\n

Compared with languages like English or Spanish, there\u2019s less available source data in Welsh for AI training. So to create a sufficiently large Welsh training dataset, the team used NVIDIA NIM<\/a> microservices for gpt-oss-120b<\/a> and DeepSeek-R1<\/a> to translate NVIDIA Nemotron open datasets<\/a> with over 30 million entries from English to Welsh.<\/p>\n

They used a GPU cluster through the NVIDIA DGX Cloud Lepton<\/a> platform and are harnessing hundreds of NVIDIA GH200 Grace Hopper Superchips<\/a> on Isambard-AI<\/a> \u2014 the U.K.\u2019s most powerful supercomputer, backed by \u00a3225 million in government investment<\/a> and based at University of Bristol \u2014 to accelerate their translation and training workloads.<\/p>\n

This new dataset supplements existing Welsh data from the team\u2019s previous efforts.<\/p>\n

Capturing Linguistic Nuances With Careful Evaluation<\/b><\/h2>\n

Bangor University, located in Gwynedd \u2014 the county with the highest percentage of Welsh speakers<\/a> \u2014 is supporting the new model\u2019s development with linguistic and cultural expertise.<\/p>\n

\"\"
Welsh translation of: \u201cThe aim is to ensure that Welsh remains a living, breathing language that continues to develop with the times.\u201d \u2014 Gruffudd Prys, Bangor University<\/figcaption><\/figure>\n

Prys, from the university\u2019s Welsh-language center, brings to the collaboration about two decades of experience with language technology for Welsh. He and his team are helping to verify the accuracy of machine-translated training data and manually translated evaluation data, as well as assess how the model handles nuances of Welsh that AI typically struggles with \u2014 such as the way consonants at the beginning of Welsh words change based on neighboring words.<\/p>\n

The model, as well as the Welsh training and evaluation datasets, are expected to be made available for enterprise and public sector use, supporting additional research, model training and application development.<\/p>\n

\u201cIt\u2019s one thing to have this AI capability exist in Welsh, but it\u2019s another to make it open and accessible for everyone,\u201d Prys said. \u201cThat subtle distinction can be the difference between this technology being used or not being used.\u201d<\/p>\n

Deploy Sovereign AI Models With NVIDIA Nemotron, NIM Microservices<\/b><\/h2>\n

The framework used to develop UK-LLM\u2019s model for Welsh can serve as a foundation for multilingual AI development around the world.<\/p>\n

Benchmark-topping Nemotron models, data and recipes are publicly available for developers to build reasoning models tailored to virtually any language, domain and workflow. Packaged as NVIDIA NIM microservices, Nemotron models are optimized for cost-effective compute and run anywhere, from laptop to cloud.<\/p>\n

Europe\u2019s enterprises will be able to run open, sovereign models on the Perplexity<\/a> AI-powered search engine.<\/p>\n

Get started with <\/i>NVIDIA Nemotron<\/i><\/a>.<\/i><\/p>\n

\n

Welsh translation:\u00a0<\/em><\/p>\n

Ymestyn Ar Draws yr Ynysoedd: Mae DU-LLM yn Dod \u00e2 Deallusrwydd Artiffisial i Ieithoedd y DU Gyda NVIDIA Nemotron<\/strong><\/h2>\n

Wedi\u2019i hyfforddi ar yr uwch gyfrifiadur Isambard-AI, mae model newydd a ddatblygwyd gan University College London, NVIDIA a Phrifysgol Bangor yn manteisio ar dechnegau a setiau data ffynhonnell agored NVIDIA Nemotron i alluogi rhesymu Deallusrwydd Artiffisial ar gyfer y Gymraeg ac ieithoedd eraill y DU ar gyfer gwasanaethau cyhoeddus gan gynnwys gofal iechyd, addysg ac adnoddau cyfreithiol.<\/em><\/p>\n

Ieithoedd Celtaidd \u2014 gan gynnwys Cernyweg, Gwyddeleg, Gaeleg yr Alban a Chymraeg \u2014 yw ieithoedd byw hynaf y DU. Er mwyn grymuso eu siaradwyr, mae menter Deallusrwydd Artiffisial sofran<\/a> y DU-LLM yn adeiladu model Deallusrwydd Artiffisial yn seiliedig ar NVIDIA Nemotron<\/a> a all resymu yn Saesneg a Chymraeg hefyd, iaith a siaredir gan tua 850,000 o bobl<\/a> yng Nghymru heddiw.<\/p>\n

Bydd galluogi rhesymu Deallusrwydd Artiffisial o ansawdd uchel yn y Gymraeg yn cefnogi\u2019r ddarpariaeth o wasanaethau cyhoeddus gan gynnwys gofal iechyd, addysg ac adnoddau cyfreithiol yn yr iaith.<\/p>\n

\u201cRwyf am i bob cwr o\u2019r DU allu harneisio manteision deallusrwydd artiffisial. Drwy alluogi deallusrwydd artiffisial i resymu yn y Gymraeg, rydym yn sicrhau bod gwasanaethau cyhoeddus \u2014 o ofal iechyd i addysg \u2014 yn hygyrch i bawb, yn yr iaith maen nhw\u2019n byw ynddi,\u201d meddai Prif Weinidog y DU, Keir Starmer. \u201cMae hon yn enghraifft bwerus o sut y gall y dechnoleg dddiweddaraf, wedi\u2019i hyfforddi ar uwch gyfrifiadur deallusrwydd artiffisial mwyaf datblygedig y DU ym Mryste, wasanaethu lles y cyhoedd, amddiffyn treftadaeth ddiwylliannol a datgloi cyfleoedd ledled y wlad.\u201d<\/p>\n

Mae prosiect DU-LLM, a sefydlwyd yn 2023 fel BritLLM ac a arweinir gan University College London, wedi rhyddhau dau fodel ar gyfer ieithoedd y DU yn flaenorol. Mae ei fodel newydd ar gyfer y Gymraeg, a ddatblygwyd mewn cydweithrediad \u00e2 Phrifysgol Bangor Cymru ac NVIDIA, yn cyd-fynd ag ymdrechion llywodraeth Cymru i hybu defnydd gweithredol o\u2019r iaith, gyda\u2019r nod o gyflawni miliwn o siaradwyr erbyn 2050 \u2014 menter o\u2019r enw Cymraeg 2050<\/a>.<\/p>\n

Bydd darparwr cwmwl Deallusrwydd Artiffisial yn y DU, Nscale, yn sicrhau bod y model newydd ar gael i ddatblygwyr trwy ei ryngwyneb rhaglennu rhaglenni (API).<\/p>\n

\u201cY nod yw sicrhau bod y Gymraeg yn parhau i fod yn iaith fyw, sy\u2019n anadlu ac sy\u2019n parhau i ddatblygu gyda\u2019r oes,\u201d meddai Gruffudd Prys, uwch derminolegydd a phennaeth yr Uned Technolegau Iaith yng Nghanolfan Bedwyr, canolfan y brifysgol ar gyfer gwasanaethau, ymchwil a thechnoleg y Gymraeg. \u201cMae deallusrwydd artiffisial yn dangos potensial aruthrol i helpu gyda chaffael y Gymraeg fel ail iaith yn ogystal \u00e2 galluogi siaradwyr brodorol i wella eu sgiliau iaith.\u201d<\/p>\n

Gallai\u2019r model newydd hwn hefyd roi hwb i hygyrchedd adnoddau Cymraeg drwy alluogi sefydliadau cyhoeddus a busnesau sy\u2019n gweithredu yng Nghymru i gyfieithu cynnwys neu ddarparu gwasanaethau sgwrsfot dwyieithog. Gall hyn helpu grwpiau gan gynnwys darparwyr gofal iechyd, addysgwyr, darlledwyr, manwerthwyr a pherchnogion bwytai i sicrhau bod eu cynnwys ysgrifenedig yr un mor hawdd ar gael yn y Gymraeg ag y mae yn Saesneg.<\/p>\n

Y tu hwnt i\u2019r Gymraeg, mae t\u00eem y DU-LLM yn anelu at gymhwyso\u2019r un fethodoleg a ddefnyddiwyd ar gyfer ei fodel newydd i ddatblygu modelau Deallusrwydd Artiffisial ar gyfer ieithoedd eraill a siaredir ledled y DU fel Cernyweg, Gwyddeleg, Sgoteg a Gaeleg yr Alban \u2014 yn ogystal \u00e2 gweithio gyda chydweithwyr rhyngwladol i adeiladu modelau ar gyfer ieithoedd o Affrica a De-ddwyrain Asia.<\/p>\n

\u201cMae\u2019r cydweithrediad hwn gydag NVIDIA a Phrifysgol Bangor wedi ein galluogi i greu data hyfforddi newydd a hyfforddi model newydd mewn amser record, gan gyflymu ein nod o adeiladu\u2019r model iaith gorau erioed ar gyfer y Gymraeg,\u201d meddai Pontus Stenetorp, yr athro prosesu iaith naturiol a dirprwy gyfarwyddwr y Ganolfan Deallusrwydd Artiffisial yn University College London. \u201cEin nod yw cymryd y mewnwelediadau a gafwyd o\u2019r model Cymraeg a\u2019u cymhwyso i ieithoedd lleiafrifol eraill, yn y DU ac ar draws y byd.<\/p>\n

Manteisio ar Seilwaith Deallusrwydd Artiffisial Sofran ar gyfer Datblygu Model<\/strong>\u00a0<\/strong><\/h2>\n

Mae\u2019r model newydd ar gyfer y Gymraeg yn seiliedig ar NVIDIA Nemotron<\/a>, teulu o fodelau ffynhonnell agored sy\u2019n cynnwys pwysau, setiau data a ryseitiau agored.Mae\u2019r t\u00eem datblygu DU-LLM wedi manteisio ar fodel 49-biliwn-paramedr Llama Nemotron Super a model 9-biliwn-paramedr Nemotron Nano, gan eu h\u00f4l hyfforddi<\/a> ar ddata iaith Gymraeg.<\/p>\n

O\u2019i gymharu ag ieithoedd fel Saesneg neu Sbaeneg, mae llai o ddata ffynhonnell ar gael yn y Gymraeg ar gyfer hyfforddiant Deallusrwydd Artiffisial. Felly, er mwyn creu set ddata hyfforddi Cymraeg ddigon mawr, defnyddiodd y t\u00eem ficrowasanaethau NVIDIA NIM<\/a> ar gyfer gpt-oss-120b<\/a> a DeepSeek-R1<\/a> i gyfieithu setiau data agored NVIDIA<\/a> gyda dros 30 miliwn o gofnodion o\u2019r Saesneg i\u2019r Gymraeg.<\/p>\n

Defnyddion nhw glwstwr GPU drwy blatfform NVIDIA DGX Cloud Lepton<\/a> ac yn harneisio cannoedd o Uwchsglodion NVIDIA GH200 Grace Hopper<\/a> ar Isambard-AI<\/a> \u2014 uwchgyfrifiadur mwyaf pwerus y DU, gyda chefnogaeth \u00a3225 miliwn o fuddsoddiad gan y llywodraeth<\/a> ac wedi\u2019i leoli ym Mhrifysgol Bryste \u2014 i gyflymu eu llwythi gwaith cyfieithu a hyfforddi.<\/p>\n

Mae\u2019r set ddata newydd hon yn ategu data presennol yr iaith Gymraeg o ymdrechion blaenorol y t\u00eem.<\/p>\n

Cipio Naws Ieithyddol Gyda Gwerthusiad Gofalus<\/strong><\/h2>\n

Mae Prifysgol Bangor, sydd wedi\u2019i lleoli yng Ngwynedd \u2014 y sir gyda\u2019r ganran uchaf o siaradwyr Cymraegs<\/a> \u2014 yn cefnogi datblygiad y model newydd gydag arbenigedd ieithyddol a diwylliannol.<\/p>\n

Mae Prys, o ganolfan Gymraeg y brifysgol, yn dod \u00e2 thua dau ddegawd o brofiad gyda thechnoleg iaith ar gyfer y Gymraeg i\u2019r cydweithrediad. Mae ef a\u2019i d\u00eem yn helpu i wirio cywirdeb data hyfforddi a gyfieithir gan beiriannau a data gwerthuso a gyfieithir \u00e2 llaw, yn ogystal ag asesu sut mae\u2019r model yn ymdrin \u00e2 naws Gymraeg y mae Deallusrwydd Artiffisial fel arfer yn cael trafferth \u00e2 nhw \u2014 megis y ffordd y mae cytseiniaid ar ddechrau geiriau Cymraeg yn newid yn seiliedig ar eiriau cyfagos.<\/p>\n

Disgwylir i\u2019r model, yn ogystal \u00e2\u2019r setiau data hyfforddiant a gwerthuso\u2019r Gymraeg, fod ar gael i fentrau a\u2019r sector cyhoeddus eu defnyddio, gan gefnogi ymchwil ychwanegol, hyfforddiant modelu a datblygu rhaglenni.<\/p>\n

\u201cMae\u2019n un peth cael y gallu Deallusrwydd Artiffisial hwn yn bodoli yn y Gymraeg, ond mae\u2019n beth arall ei wneud yn agored ac yn hygyrch i bawb,\u201d meddai Prys. \u201cGall y gwahaniaeth cynnil hwnnw fod y gwahaniaeth rhwng y dechnoleg hon yn cael ei defnyddio ai peidio.\u201d<\/p>\n

Defnyddio Modelau Deallusrwydd Artiffisial Sofran Gyda NVIDIA Nemotron, Microwasanaethau NIM<\/strong><\/h2>\n

Gall y fframwaith a ddefnyddiwyd i ddatblygu model DU-LLM ar gyfer y Gymraeg fod yn sylfaen ar gyfer datblygu Deallusrwydd Artiffisial amlieithog ledled y byd.<\/p>\n

Mae modelau, data a ryseitiau Nemotron, sy\u2019n cyrraedd y brig, ar gael yn gyhoeddus i ddatblygwyr er mwyn iddynt adeiladu modelau rhesymu sydd wedi\u2019u teilwra i bron unrhyw iaith, parth a llif gwaith. Wedi\u2019u pecynnu fel microgwasanaethau NVIDIA NIM, mae modelau Nemotron wedi\u2019u hoptimeiddio ar gyfer cyfrifiadura cost-effeithiol a rhedeg yn unrhyw le, o liniadur i\u2019r cwmwl.<\/p>\n

Bydd mentrau Ewrop yn gallu rhedeg modelau agored, sofran ar y peiriant chwilio Perplexity<\/a> wedi\u2019i bweru gan Ddeallusrwydd Artiffisial.<\/p>\n

Dewch i ddechrau arni gyda <\/em>NVIDIA Nemotron<\/em><\/a>.<\/em><\/p>\n<\/p><\/div>\n","protected":false},"excerpt":{"rendered":"

https:\/\/blogs.nvidia.com\/blog\/uk-llm-nemotron\/<\/p>\n","protected":false},"author":0,"featured_media":4266,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[3],"tags":[],"_links":{"self":[{"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/posts\/4265"}],"collection":[{"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/comments?post=4265"}],"version-history":[{"count":0,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/posts\/4265\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/media\/4266"}],"wp:attachment":[{"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/media?parent=4265"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/categories?post=4265"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/tags?post=4265"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}