HejuruUbuyobozi bukurikira
Ibiranga AI Ibiranga kurubuga rusange
Sosiyete
AMABWIRIZA Yibanze
Label smoothing replaces hard one-hot targets with a softened target distribution during training.
Umva iki gitaboAI-byavuzwe · ~2 min
It can regularize some models, but its effects on accuracy, calibration, and distillation depend on the task, model, and evaluation.
With K classes, one common label-smoothing convention assigns 1 − ε to the correct class and distributes ε across the remaining classes; implementations may instead distribute ε across all classes, so check the loss definition. This prevents the target from demanding probability one for the labeled class and can discourage extreme logits. The Inception-v3 work and the Transformer paper are prominent documented uses of the method. Research has found calibration and generalization benefits in some settings, while also showing that smoothing can remove information about similarities among incorrect classes and weaken a model as a knowledge-distillation teacher. It is therefore not a universal accuracy or calibration improvement. Choose ε and evaluate calibration, task quality, and downstream distillation on representative validation data.
IshushoNet itondekanya: Inception-v3 yakoresheje label yoroshye (epsilon 0.1) kugirango izamure top-1 neza kandi igabanye kwigirira icyizere.
Ubusobanuro bwimashini: umwimerere wa Transformer washyizeho label yoroshye ya 0.1, ucuruza gato urujijo kumanota menshi ya BLEU.
Kumenyekanisha imvugo: intego zoroheje zigabanya kutizerana birenze urugero no kunoza kalibrasi kumajwi asakuza.
Uburyo bwo gufata amashusho yubuvuzi: koroshya gutanga umusaruro mwiza-wahinduwe neza, ingenzi mugihe amanota yicyizere amenyesha ibyemezo byubuvuzi.
Ikirango cyoroheje gikomeza kuba indashyikirwa mumahugurwa manini, ariko ubushakashatsi bugenda bwerekeza kumihindagurikire kandi yize yoroshye ihindura epsilon kurugero cyangwa ishuri aho gukoresha agaciro kamwe. Uburyo bwibanze bwa Calibration nko gutakaza kwibanda hamwe no gupima ubushyuhe akenshi bipimwa cyangwa bigahuzwa nabyo. Mugihe icyitegererezo gikura kandi ikigereranyo kidashidikanywaho kigereranywa n’umutekano, tegereza koroshya kuba igikoresho kimwe muri benshi kugirango batange amanota yizewe, witondere neza amakimbirane azwi no kubeshya.
Ibaruwa y'amakuru y'ubuntu
Inkuru eshatu zizewe za AI buri gitondo cy'iminsi y'akazi, zanditse mu rurimi rworoshye. Ubuntu iteka ryose, nta kwamamaza.
Imeri imwe buri munsi y'akazi. Kwiyimura mu buryo bumwe. Ntitwigera tugurisha cyangwa ngo dusangire aderesi yawe.
Isuzume
Ibisubizo ako kanya kuri buri kibazo, n'icyemezo gishobora gusangirwa gifite ID yagenzurwa umaze gutsinda isomo.
Fasha uburezi bwa AI bw'ubuntu. AI Understanding ni umuryango utegamiye kuri leta wa 501(c)(3) — nta kwamamaza, nta kwishyura, iteka ryose. Tanga impano
Label smoothing replaces hard one-hot targets with a softened target distribution during training. It can regularize some models, but its effects on accuracy, calibration, and distillation depend on the task, model, and evaluation.
Aho kugirango intego igoye 1.0 / 0.0, ikirango cyorohereza umukoro (1 - epsilon) kumurongo wukuri kandi ikwirakwiza epsilon mubindi.
Smoothing softens the target distribution so the training objective does not demand a one-hot probability vector.
Inception-v3 and the Transformer paper are prominent documented uses of label smoothing; this wording makes no claim that either originated the technique.
Research reports calibration effects that depend on the model and evaluation setting; label smoothing is not a universal calibration guarantee.
Mugukomeza cluster no gusibanganya umubano woroshye hagati yamasomo, koroshya birashobora gukuraho amakuru gusibanganya ubumenyi bishingiye.
Komeza wige
Abandi bayobozi batoranijwe kuriyi ngingo
HejuruUbuyobozi bukurikira
Ibiranga AI Ibiranga kurubuga rusange
Sosiyete