Search results

Merging models without access to training data

Status: Open.

#10 In bloomberg/dataless-model-merging;

· MoeinSorkhei opened on Mar 22, 2024

Gram matrix for non-masked-out input activations only

Status: Open.

#9 In bloomberg/dataless-model-merging;

· dptam opened on Feb 15, 2024

Datasets Issues for Reproduce the Experiment

Status: Open.

#8 In bloomberg/dataless-model-merging;

· NoraLiu525 opened on Aug 9, 2023

Activation-based matching code

Status: Open.

#7 In bloomberg/dataless-model-merging;

· XueY-L opened on Aug 4, 2023

How to implement RegMean for GPT-like model?

Status: Open.

#4 In bloomberg/dataless-model-merging;

· A11en0 opened on Jul 6, 2023