GitLab

This reduces the built-in linux version to not use any vector extensions
which enables the resulting builds to run under Rosetta on MacOS in
Docker.  Then at runtime it checks for the actual CPU vector
extensions and loads the best CPU library available

In some cases we may want multiple variants for a given GPU type or CPU.
This adds logic to have an optional Variant which we can use to select
an optimal library, but also allows us to try multiple variants in case
some fail to load.

This can be useful for scenarios such as ROCm v5 vs v6 incompatibility
or potentially CPU features.

* increase minimum cuda overhead and fix minimum overhead for multi-gpu

* fix multi gpu overhead

* limit overhead to 10% of all gpus

* better wording

* allocate fixed amount before layers

* fixed only includes graph alloc

* select layers based on estimated model memory usage

* always account for scratch vram

* dont load +1 layers

* better estmation for graph alloc

* Update gpu/gpu_darwin.go
Co-authored-by: Bruce MacDonald <brucewmacdonald@gmail.com>

* Update llm/llm.go
Co-authored-by: Bruce MacDonald <brucewmacdonald@gmail.com>

* Update llm/llm.go

* add overhead for cuda memory

* Update llm/llm.go
Co-authored-by: Bruce MacDonald <brucewmacdonald@gmail.com>

* fix build error on linux

* address comments

---------
Co-authored-by: Bruce MacDonald <brucewmacdonald@gmail.com>

On linux, we link the CPU library in to the Go app and fall back to it
when no GPU match is found. On windows we do not link in the CPU library
so that we can better control our dependencies for the CLI.  This fixes
the logic so we correctly fallback to the dynamic CPU library
on windows.

This switches the default llama.cpp to be CPU based, and builds the GPU variants
as dynamically loaded libraries which we can select at runtime.

This also bumps the ROCm library to version 6 given 5.7 builds don't work
on the latest ROCm library that just shipped.

This allows the CPU only builds to work on systems with Radeon cards

Run the server.cpp directly inside the Go runtime via cgo
while retaining the LLM Go abstractions.

- remove ggml runner
- automatically pull gguf models when ggml detected
- tell users to update to gguf in the case automatic pull fails
Co-Authored-By: Jeffrey Morgan <jmorganca@gmail.com>

This reverts commit 7a0899d6.

- update chat docs
- add messages chat endpoint
- remove deprecated context and template generate parameters from docs
- context and template are still supported for the time being and will continue to work as expected
- add partial response to chat history

* add `"format": "json"` as an API parameter
---------
Co-authored-by: Bruce MacDonald <brucewmacdonald@gmail.com>

add error for falcon and starcoder vocab compatibility
---------
Co-authored-by: Bruce MacDonald <brucewmacdonald@gmail.com>

---------
Co-authored-by: Michael Yang <mxyng@pm.me>

* remove tmp directories created by previous servers

* clean up on server stop

* Update routes.go

* Update server/routes.go
Co-authored-by: Jeffrey Morgan <jmorganca@gmail.com>

* create top-level temp ollama dir

* check file exists before creating

---------
Co-authored-by: Jeffrey Morgan <jmorganca@gmail.com>
Co-authored-by: Michael Yang <mxyng@pm.me>

get model and file type from bin file

* remove c code
* pack llama.cpp
* use request context for llama_cpp
* let llama_cpp decide the number of threads to use
* stop llama runner when app stops
* remove sample count and duration metrics
* use go generate to get libraries
* tmp dir for running llm

warning F16 uses significantly more memory than quantized model so the
standard requires don't apply.