List your company
ResearchFormerly WizWand

Class-Incremental Learning on CIFAR

89.6AccuracyRanPAC
-2.54421.37845.369.222Feb 13, 2024
Updated 8mo ago

Evaluation results

MethodLinks
RanPACBackbone=ViT-B/16, Pre-training=ImageNet-21K, Task Setting=2 init classes, 1 class per task2024.0289.6
RanPAC-impBackbone=ViT-B/16, Pre-training=ImageNet-21K, Task Setting=2 init classes, 1 class per task2024.0289.4
RanPACBackbone=ViT-B/16, Pre-training=ImageNet-1K, Task Setting=2 init classes, 1 class per task2024.0288.2
RanPAC-impBackbone=ViT-B/16, Pre-training=ImageNet-1K, Task Setting=2 init classes, 1 class per task2024.0287.8
FosterBackbone=ViT-B/16, Pre-training=ImageNet-21K, Task Setting=2 init classes, 1 class per task2024.0287.3
SLCABackbone=ViT-B/16, Pre-training=ImageNet-21K, Task Setting=2 init classes, 1 class per task2024.0286.8
RanDumbBackbone=ViT-B/16, Pre-training=ImageNet-21K, Task Setting=2 init classes, 1 class per task2024.0286.8
SLCABackbone=ViT-B/16, Pre-training=ImageNet-1K, Task Setting=2 init classes, 1 class per task2024.0286.3
RanDumbBackbone=ViT-B/16, Pre-training=ImageNet-1K, Task Setting=2 init classes, 1 class per task2024.0284.5
NCMBackbone=ViT-B/16, Pre-training=ImageNet-21K, Task Setting=2 init classes, 1 class per task2024.0283.5
Adam-SSFBackbone=ViT-B/16, Pre-training=ImageNet-21K, Task Setting=2 init classes, 1 class per task2024.0282.7
Adam-AdaptBackbone=ViT-B/16, Pre-training=ImageNet-21K, Task Setting=2 init classes, 1 class per task2024.0282.4
Adam-VPTBackbone=ViT-B/16, Pre-training=ImageNet-1K, Task Setting=2 init classes, 1 class per task2024.0279.3
NCMBackbone=ViT-B/16, Pre-training=ImageNet-1K, Task Setting=2 init classes, 1 class per task2024.0278.3
Adam-AdaptBackbone=ViT-B/16, Pre-training=ImageNet-1K, Task Setting=2 init classes, 1 class per task2024.0276.7
Adam-SSFBackbone=ViT-B/16, Pre-training=ImageNet-1K, Task Setting=2 init classes, 1 class per task2024.0276
Adam-FTBackbone=ViT-B/16, Pre-training=ImageNet-1K, Task Setting=2 init classes, 1 class per task2024.0272.6
iCARLBackbone=ViT-B/16, Pre-training=ImageNet-1K, Task Setting=2 init classes, 1 class per task2024.0272.4
iCARLBackbone=ViT-B/16, Pre-training=ImageNet-21K, Task Setting=2 init classes, 1 class per task2024.0271.6
Adam-VPTBackbone=ViT-B/16, Pre-training=ImageNet-21K, Task Setting=2 init classes, 1 class per task2024.0270.8
MemoBackbone=ViT-B/16, Pre-training=ImageNet-1K, Task Setting=2 init classes, 1 class per task2024.0269.8
Adam-FTBackbone=ViT-B/16, Pre-training=ImageNet-21K, Task Setting=2 init classes, 1 class per task2024.0265.7
FosterBackbone=ViT-B/16, Pre-training=ImageNet-1K, Task Setting=2 init classes, 1 class per task2024.0252.2
FinetuneBackbone=ViT-B/16, Pre-training=ImageNet-21K, Task Setting=2 init classes, 1 class per task2024.022.8
CODA-PromptBackbone=ViT-B/16, Pre-training=ImageNet-1K, Task Setting=2 init classes, 1 class per task2024.022.6
L2PBackbone=ViT-B/16, Pre-training=ImageNet-1K, Task Setting=2 init classes, 1 class per task2024.022.4
DualPromptBackbone=ViT-B/16, Pre-training=ImageNet-1K, Task Setting=2 init classes, 1 class per task2024.022.3
FinetuneBackbone=ViT-B/16, Pre-training=ImageNet-1K, Task Setting=2 init classes, 1 class per task2024.021