What is Parameter-Efficient Fine-Tuning (PEFT)? Definition | AI Understanding

Tsanangudzo

Nzira dzinogadzirisa mamodheru nekudzidzisa diki diki rekuwedzera paramita.

Mashoko anoenderana

Kugadziriswa kwakanaka

Kuenderera mberi nekudzidziswa padomeine-chaiyo data kugadzirisa iyo isati yadzidziswa modhi kune rimwe basa.

LoRA (Low-Rank Adaptation)

A parameter-inoshanda zvakanaka-tuning nzira inowedzera yakaderera-rank adapter matrices.

Mushure mekudzidziswa

Matanho ekudzidzisa anoshandiswa mushure mekutanga kudzidziswa, senge kuraira tuning, optimization yekuda, uye kuchengetedza tuning.

Instruction Tuning

Kunyatsogadzirisa modhi pamirairo-mhinduro mbiri kuti uvandudze basa rinotevera.

Iyo hyperparameter yekudzidzira inodzora kuti yakawanda sei paramita inoshandura nhanho yega yega yekuvandudza.

DPO (Direct Preference Optimization)

Nzira yekudzidzisa iyo inonatsa-tuni modhi yakananga pamapeya ekuda pasina kuda mubairo wakasiyana.

Dzidza zvakawanda mumadhairekitori edu emahara

Fine-Tuning Hard Parameter Sharing in Multi-Task Networks Gumbel-Softmax and Reparameterization Rejection Sampling Fine-Tuning

Onawo

Parameter Perplexity Overfitting Pipeline Open-Source Model Precision OCR (Optical Character Recognition)Pretraining

Bhurawuza yakazara AI Glossary Ongorora ese emahara AI madhairekitori