nguyentriquan Goto Github PK

architecture-and-rehearsal-in-continual-learning

My research about Architecture based and Rehearsal based Continual Learning

continual-learning-baselines

Implement some Continual Learning approaches and experiments.

cpg-baseline

run more experiments with approach Compacting, Picking and Growing for Unforgetting Continual Learning

d3pm

Discrete Denoising Diffusion Probabilistic Models (D3PM)

dae

Class Incremental Learning with Dynamic architecture and knowledge ensemble. Utilize over-parameterization, pruning, ensemble and calibration to overcome the challenges of Class Incremental Learning

diffu-dat

Official Implementation for the ICLR2023 paper "Fuzzy Alignments in Directed Acyclic Graph for Non-autoregressive Machine Translation"

diffusionlms

Diffusion models for text generation

dst

dynamic sparse training baselines with Node Path balancing

gpm-supcon

Combination of Gradient Projection Memory and Supervised Contrastive learning for Class Incremental Learning

pruned-dnns

Generalization in Pruned Deep Neural Networks. Measure the lipschitz constant of the pruned networks. Normalize the distribution of weight to improve generalization

sccl

Over-parameterized expansion and Structure Compression for Continual Learning

sccl-code

supsup-baseline

run Supermasks in Superposition