Daniel Hiltgen 58d95cc9bd Switch back to subprocessing for llama.cpp ...

This should resolve a number of memory leak and stability defects by allowing
us to isolate llama.cpp in a separate process and shutdown when idle, and
gracefully restart if it has problems.  This also serves as a first step to be
able to run multiple copies to support multiple models concurrently.

2024-04-01 16:48:18 -07:00

..

ext_server

Switch back to subprocessing for llama.cpp

2024-04-01 16:48:18 -07:00

generate

Switch back to subprocessing for llama.cpp

2024-04-01 16:48:18 -07:00

llama.cpp @ ad3a0505e3

Bump llama.cpp to b2527

2024-03-25 13:47:44 -07:00

patches

Bump llama.cpp to b2474

2024-03-23 09:54:56 +01:00

ggla.go

refactor model parsing

2024-04-01 13:16:15 -07:00

ggml.go

update memory calcualtions

2024-04-01 13:16:32 -07:00

gguf.go

refactor model parsing

2024-04-01 13:16:15 -07:00

llm_darwin_amd64.go

Switch back to subprocessing for llama.cpp

2024-04-01 16:48:18 -07:00

llm_darwin_arm64.go

Switch back to subprocessing for llama.cpp

2024-04-01 16:48:18 -07:00

llm_linux.go

Switch back to subprocessing for llama.cpp

2024-04-01 16:48:18 -07:00

llm_windows.go

Switch back to subprocessing for llama.cpp

2024-04-01 16:48:18 -07:00

llm.go

Switch back to subprocessing for llama.cpp

2024-04-01 16:48:18 -07:00

payload.go

Switch back to subprocessing for llama.cpp

2024-04-01 16:48:18 -07:00

server.go

Switch back to subprocessing for llama.cpp

2024-04-01 16:48:18 -07:00

status.go

Switch back to subprocessing for llama.cpp

2024-04-01 16:48:18 -07:00