Skip to content

b3078

Compare
Choose a tag to compare
@github-actions github-actions released this 04 Jun 07:37
bde7cd3
llama : offload to RPC in addition to other backends (#7640)

* llama : offload to RPC in addition to other backends

* - fix copy_tensor being called on the src buffer instead of the dst buffer

- always initialize views in the view_src buffer

- add RPC backend to Makefile build

- add endpoint to all RPC object names

* add rpc-server to Makefile

* Update llama.cpp

Co-authored-by: slaren <[email protected]>

---------

Co-authored-by: slaren <[email protected]>