Improving Fine-tuning Pre-trained Models on Small Source Code Datasets via Variational Information Bottleneck.

Jiaxing Liu Chaofeng Sha Xin Peng

Published in: SANER (2023)

Keyphrases

source code
fine tuning
open source
information bottleneck
automated tools
pre trained
software systems
information theoretic
data sets
prior knowledge
high level
support vector
small number