LoRA na PEFT
LoRA (Low-Rank Adaptation), iliyoanzishwa na Hu et al. mwaka 2022, na familia pana ya mbinu za urekebishaji wa vigezo kwa ufanisi (PEFT) hubadilisha miundo mikuu ya lugha iliyofunzwa awali kwa ajili ya kazi mpya kwa kufunza idadi ndogo tu ya vigezo vya ziada badala ya kila uzito katika modeli. Hii huwezesha urekebishaji kwa kumbukumbu ndogo sana ya GPU na hesabu huku ikiacha modeli asili ikiwa haijaguswa.
Soma mbinu kamili
Ingia kwa akaunti ya bure ili kusoma sehemu hii.
Method map
The neighbourhood of related methods — select a node to explore.
Vyanzo
- Hu, E. J. et al. (2022). LoRA: Low-Rank Adaptation of Large Language Models. ICLR. link ↗
- Lester, B. et al. (2021). The Power of Scale for Parameter-Efficient Prompt Tuning. EMNLP. DOI: 10.18653/v1/2021.emnlp-main.243 ↗
Jinsi ya kunukuu ukurasa huu
ScholarGate. (2026, June 1). Low-Rank Adaptation and Parameter-Efficient Fine-Tuning. ScholarGate. https://scholargate.app/sw/deep-learning/lora-peft
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Mtandao wa Kushawishi unaozalisha (Generative Adversarial Network - GAN)Ujifunzaji wa Kina↔ compare
- Msitu NasibuUjifunzaji wa Mashine↔ compare
- Variational AutoencoderUjifunzaji wa Kina↔ compare
- Transformer wa MaonoUjifunzaji wa Kina↔ compare
- XGBoostUjifunzaji wa Mashine↔ compare
Imerejelewa na
Umeona tatizo kwenye ukurasa huu? Ripoti au pendekeza marekebisho →