mammoth
mammoth copied to clipboard
aimagelab
An Extendible (General) Continual Learning Framework based on Pytorch - official codebase of Dark Experience for General Continual Learning
Hi, thank you for the amazing paper and repository. I am trying to run some experiments on CIFAR100 in task incremental setting with 10 different 10-way classification tasks. Can you...
Hi! I can't really surpass der, der++ performance by using xder on cifar10 do you happen to have tuned xder for Cifar10? Thanks!
Hi, I was using the library and it seems like the iCARL method is not working. I figured out the reason and it seems like while selecting the samples to...
## Main changelog: ### Improved reproducibility - validated *all* models - validated *all* datasets - introduced *tests* for all datasets and models (all pass) ### WiKi! - added *WiKi* with...
(base) root@marksheng-jgdgkqoauw1k-main:/openbayes/home/mammoth# python ./utils/main.py --model pnn --dataset seq-cifar100 --lr 0.001 Warning: GEM and A-GEM cannot be used on Windows (quadprog required) Namespace(batch_size=32, conf_host='marksheng-jgdgkqoauw1k-main', conf_jobnum='38cff68f-3a64-41c9-a4e7-be660f81e47d', conf_timestamp='2023-06-16 19:04:08.766651', dataset='seq-cifar100', debug_mode=0, disable_log=0, distributed='no',...
https://github.com/aimagelab/mammoth/blob/f3effb9dc2098c0463d214c4f8e5e69f38517e13/models/gem.py#L126 Dear authors, This part where you add data to buffer in GEM is probably wrong because size of task_labels should be equal to that of examples, but not 'samples_per_task'....
