Subira ku makuru
Guhanga udushyaAI Understanding ibisobanuro

Abashakashatsi batangiza icyitegererezo cyibanze cyo gutunganya inyandiko ya HTML yo muri Tchèque

Abashakashatsi berekanye icyitegererezo cyibanze cyitwa HTML-LM, cyagenewe gutunganya inyandiko za HTML zo muri Tchèque. Icyitegererezo gifite miriyoni 154 kandi cyahuguwe kuri miliyoni 100 zurubuga ukoresheje intego nyinshi.

4 min readRead the primary source
Source-page capture accompanying Researchers introduce compact foundation model for Czech HTML documents processing
Inyandiko y'ibanzeInkomoko yanditse
Umwanditsi
arxiv.org
Ihuza ry'inkomoko
arxiv.orghttps://arxiv.org/abs/2609.18494
Ubwoko bw'inkomoko
Inyandiko y'ibanze - itangazo ryemewe, impapuro, dosiye, cyangwa urupapuro rwambere-dusoma mu buryo butaziguye.
ImirongoSobanukirwa ibi mumasegonda 60

Tangira hano

Amagambo y'ingenzi

Icyitegererezo
Icyitegererezo kinini cyabanje gutozwa gishobora guhuzwa nimirimo myinshi yo hasi.
Ibyiciro
Igikorwa aho icyitegererezo gitanga ibitekerezo kumurongo umwe cyangwa byinshi byateganijwe mbere.
Dataset
Icyegeranyo cyingero zubatswe cyangwa zitubatswe zikoreshwa mumahugurwa, kwemeza, cyangwa kugerageza.
IsuzumeAI ni iki? Ikibazo

Byagenze bite

Abashakashatsi bakoze icyitegererezo cyibanze cyitwa HTML-LM, cyagenewe gutunganya inyandiko za HTML zo muri Tchèque. Icyitegererezo gifite miriyoni 154 kandi cyahuguwe kuri miliyoni 100 zurubuga ukoresheje intego nyinshi. Ishiraho uburyo bushya-bugezweho bwo gutondekanya no gusubira inyuma muri interineti ya Ceki.

Icyitegererezo gifite miriyoni 154 kandi cyahuguwe kuri miliyoni 100 zurubuga ukoresheje intego nyinshi.

Ishiraho uburyo bushya-bugezweho bwo gutondekanya no gusubira inyuma muri interineti ya Ceki.

Ibisobanuro birambuye: arxiv.org ↗

Impamvu ari ngombwa

Iterambere rya HTML-LM ni ingirakamaro kuko rikemura imbogamizi zuburyo buriho bwo gutunganya inyandiko zurubuga. Nicyitegererezo cyoroheje gikora kandi cyubukungu, bigatuma gikwiranye ninganda nyinshi zinganda.

Icyitegererezo cyatojwe kuri nini yinyandiko zurubuga, zemerera kwiga amakuru yimiterere arangwa muri HTML.

Ikoresha ModernBERT ishingiye ku myubakire, iyifasha gutunganya page-nyayo y'urubuga neza.

Icyitegererezo gikoreshwa mubikorwa, gutunganya ibihumbi byurubuga rwisegonda kumasegonda, bikabera igisubizo gifatika kubidukikije.

Irekurwa kubaturage munsi ya CC BY-NC 4.0, bigatuma ikoreshwa nabandi.

Icyitegererezo gishyiraho uburyo bushya-bugezweho bwo gutondekanya no gusubira inyuma muri porogaramu ya interineti ya Tchèque, irenga kodegisi nini nini na LLM nto.

Interactive Mechanism

Uburyo bukoreshwa: Uburyo bukora

Shakisha ikoranabuhanga ryihishe inyuma yiri terambere.

Thinking Budget (Test-Time Tokens):1,024 tokens
Complex Accuracy79%Math & Code Logic
Latency3.2sTime to first full output
Inference Cost$0.0092Per query estimated
Reasoning StyleStep VerificationInternal chain depth
Active Thinking Trace:
1Deconstruct user problem into formal constraints
2Propose candidate hypotheses & step-by-step calculation
3Self-correction: Backtrack and refute subtle edge cases
4Exhaustive consistency check & final output synthesis
Core takeaway: Test-time compute fundamentally changes AI economics. Instead of only scaling during pre-training, giving reasoning models more tokens at inference time allows them to systematically solve PhD-level STEM problems.
Kugenzura Ibitekerezo Byagenzuwe+10 Points
What is AI? Quiz

A route planner searches possible journeys using explicit rules. What does this illustrate about AI?

Ibyo kureba

Iterambere rya HTML-LM nintambwe yingenzi mugushiraho isi yose, yujuje ubuziranenge bwerekana inyandiko zurubuga ahantu h’inganda nyinshi.

Imikorere yicyitegererezo nubushobozi bwubukungu bituma iba igisubizo cyiza kubidukikije.

Ubushobozi bwikitegererezo bwo gutunganya page-nyayo y'urubuga rwose bituma iba igisubizo gifatika kumurongo mugari wa porogaramu.

Icyitegererezo cyo gusohora kubaturage munsi ya CC BY-NC 4.0 ituma ikoreshwa nabandi.

Icyitegererezo cyoherejwe mubikorwa kigaragaza ubushobozi bwacyo bwo gutunganya inyandiko ibihumbi nibihumbi kumasegonda.

Imikorere yicyitegererezo muri domaine ya enterineti ya Tchèque ishyiraho uburyo bushya bugezweho, burenga kodegisi nini nini na LLM nto.

Ibijyanye nuyobora & ibibazo

AI ni iki?Moderi ya AI YasobanuweAbahinduraAmahugurwa ya AIGerageza ibyo uzi - gerageza ikibazo cya AI kubuntuReba ijambo AI mumagambo yacuKurikiza icyerekezo cya AI cyo kurekura
Basanze ari ingirakamaro?