GitLab

On the smaller GPUs, the initial model load of llama2 took over 30s (the
default timeout for the DoGenerate helper)

Workaround broken ROCm p2p copy

* stop token parsing fix

* add stop test

Statically link c++ and thread lib on windows

This makes sure we statically link the c++ and thread library on windows
to avoid unnecessary runtime dependencies on non-standard DLLs

update named templates

update message processing

* server: fix unneeded model reloads when setting `OLLAMA_NUM_PARALLEL`

* remove whitespace change

* undo some changes

Enable the build flag for llama.cpp to use CPU copy for multi-GPU scenarios.

llm: remove ambiguous comment when putting upper limit on predictions to avoid infinite generation (#5535)

This reverts commit 5796bfc4.

* Revert "fix cmake build (#5505)"

This reverts commit 4fd5f352.

* llm: fix missing dylibs by restoring old build behavior

* crlf -> lf

* llm: put back old include dir

* llm: update link paths for old submodule commits

Always go build in CI generate steps

With the recent cgo changes, bugs can sneak through
if we don't make sure to `go build` all the permutations

This was leading to ambiguity and confusion in ollama.com, and is not
used anywhere in ollama at the moment. Once manifests are addressable by
digest, we can add this back in, and in a way that is more tailored to
the concept of addressing a manifest by digest.

ensure runtime model changes (template, system prompt, messages,
options) are captured on model updates without needing to reload the
server

* Use common prefix to select slot

* actually report `longest`

Prevent loading models larger than total memory

* Fix assert on small embedding inputs

* Update llm/patches/09-pooling.diff