GitHub tau qhib kev txhim kho hauv kev siv tshuab kev kawm rau kev tshawb nrhiav thiab kev tshuaj xyuas

GitHub tswvcuab peb tes num CodeSearchNet, uas tau npaj cov qauv kev kawm tshuab thiab cov ntaub ntawv tsim nyog rau kev txheeb xyuas, faib thiab txheeb xyuas cov lej hauv ntau hom lus programming. CodeSearchNet, zoo ib yam li ImageNet, suav nrog cov ntawv sau loj ntawm cov lej cim nrog cov lus piav qhia uas ua raws li txoj cai ua. Cov khoom siv rau kev cob qhia cov qauv thiab cov piv txwv ntawm kev siv CodeSearchNet tau sau rau hauv Python siv Tensorflow lub moj khaum thiab faib los ntawm nyob rau hauv MIT daim ntawv tso cai.

Thaum tsim CodeSearchNet, cov ntawv nyeem cov lus piav qhia cov thev naus laus zis tau siv, ua rau lub tshuab kev kawm kom coj mus rau hauv tus account tsis yog cov yam ntxwv sib xws nkaus xwb, tab sis kuj yog lub ntsiab lus ntawm kev ua los ntawm cov cai. GitHub system siv rau hauv kev sim ntawm kev teeb tsa semantic code tshawb nrhiav siv cov lus nug ntawm yam lus (piv txwv li, thaum thov "tso npe ntawm cov hlua", code nrog rau kev siv ntawm cov algorithms sib raug tau tshwm sim).

Cov ntaub ntawv uas tau npaj tseg suav nrog ntau dua 2 lab cov kab lus sib txuas lus, npaj raws li cov ntawv sau ntawm cov tsev qiv ntawv qhib uas twb muaj lawm. Txoj cai npog tag nrho cov ntawv nyeem ntawm tus kheej lub luag haujlwm lossis cov txheej txheem, thiab cov lus piav qhia txog kev ua haujlwm los ntawm kev ua haujlwm (cov ntaub ntawv nthuav dav tau muab). Tam sim no, datasets tau npaj rau Python, JavaScript, Ruby, Go, Java thiab PHP. Piv txwv yog muab los ntawm kev siv cov ntaub ntawv thov rau kev cob qhia ntau hom neural networks, suav nrog Neural-bag-Ob-Words, RNN, Kev Taw Qhia Tus Kheej (BERT) thiab 1D-CNN + Self-Attention Hybrid.

Txhawm rau tsim cov txheej txheem tshawb nrhiav lus ntuj, ib txheej ntawm CodeSearchNet Challenge tau npaj ntxiv, suav nrog
99 ua dv Cov lus nug nrog txog 4 txhiab tus kws tshaj lij cov lus piav qhia piav qhia qhov feem ntau yuav raug khi rau hauv CodeSearchNet Corpus dataset, suav txog 6 lab txoj hauv kev thiab kev ua haujlwm (teeb loj txog 20 GB). Qhov Kev Sib Tw CodeSearchNet tuaj yeem ua tus qauv ntsuas rau kev ntsuas qhov ua tau zoo ntawm qee txoj hauv kev rau kev tshawb nrhiav cov lus ntuj. Siv cov cuab yeej KubeFlow npaj
Piv txwv code nrhiav cav.

Tau qhov twg los: opennet.ru

Ntxiv ib saib