Merge Models with Non-Standard Architectures (e.g., Multimodal Models)#450

Closed

ElliotStein wants to merge 24 commits intomainfrom architecture-agnostic

+782-52

Commits on Oct 18, 2024

Generalise mergekit to work for any architecture: implemented AutomaticArchitectureInfo and get_model_parameter_names
ElliotStein
committed

Commits on Oct 23, 2024

Commits on Oct 25, 2024

Merge branch 'main' into architecture-agnostic
ElliotStein
committed

Commits on Oct 31, 2024

Commits on Nov 5, 2024

Commits on Nov 19, 2024

script to manually add weights from base model after merging submodule. And minor fixes.
ElliotStein
committed

Commits on Dec 2, 2024

Address PR feedback: resolve merge architecture error
ElliotStein
committed

Commits on Dec 6, 2024

Improve robustness and logging in architecture functions
ElliotStein
committed

Commits on Dec 9, 2024