GitLab

* Fix assert on small embedding inputs

* Update llm/patches/09-pooling.diff

* openai compatibility

* Revert "openai compatibility"

This reverts commit d3f98a811e00fc497d889c8c45b0cfec5b64690c.

* remove erroneous subtraction of prompt cache

When ollama is running a long time, tmp cleaners can remove the
runners.  This tightens up a few corner cases on arm macs where
we failed with "server cpu not listed in available servers map[]"

On windows, if the model dir contained unicode characters
clip models would fail to load.  This fixes the file name
handling in clip.cpp to support utf16 on windows.

This uses nil as undefined for a cleaner implementation.

* Do not shift context for sliding window models

* truncate prompt > 2/3 tokens

* only target gemma2

Previously, some costly things were causing the loading of GGUF files
and their metadata and tensor information to be VERY slow:

  * Too many allocations when decoding strings
  * Hitting disk for each read of each key and value, resulting in a
    not-okay amount of syscalls/disk I/O.

The show API is now down to 33ms from 800ms+ for llama3 on a macbook pro
m3.

This commit also prevents collecting large arrays of values when
decoding GGUFs (if desired). When such keys are encountered, their
values are null, and are encoded as such in JSON.

Also, this fixes a broken test that was not encoding valid GGUF.

This adjusts our default settings to enable multiple models and parallel
requests to a single model.  Users can still override these by the same
env var settings as before.  Parallel has a direct impact on
num_ctx, which in turn can have a significant impact on small VRAM GPUs
so this change also refines the algorithm so that when parallel is not
explicitly set by the user, we try to find a reasonable default that fits
the model on their GPU(s).  As before, multiple models will only load
concurrently if they fully fit in VRAM.

If we try to use mmap when the model is larger than the system free space, loading is slower than the no-mmap approach.

The recent refactoring of the memory prediction assumed all layers
are the same size, but for some models (like deepseek-coder-v2) this
is not the case, so our predictions were significantly off.

Prior to this change, we logged the memory prediction multiple times
as the scheduler iterates to find a suitable configuration, which can be
confusing since only the last log before the server starts is actually valid.
This now logs once just before starting the server on the final configuration.
It also reports what library instead of always saying "offloading to gpu" when
using CPU.

On Windows, recent llama.cpp changes make mmap slower in most
cases, so default to off.  This also implements a tri-state for
use_mmap so we can detect the difference between a user provided
value of true/false, or unspecified.

nvcc supports parallelism (threads) and cmake + make can use -j,
while msbuild requires /p:CL_MPcount=8

This reverts commit 0577af98.

We update the PATH on windows to get the CLI mapped, but this has
an unintended side effect of causing other apps that may use our bundled
DLLs to get terminated when we upgrade.

* llm: update llama.cpp submodule to `7c26775`

* disable `LLAMA_BLAS` for now

* `-DLLAMA_OPENMP=off`

Make the build faster

Implement support for GPU env var workarounds, and leverage
this for the Vega RX 56 which needs
HSA_ENABLE_SDMA=0 set to work properly

Still not complete, needs some refinement to our prediction to understand the
discrete GPUs available space so we can see how many layers fit in each one
since we can't split one layer across multiple GPUs we can't treat free space
as one logical block

This reverts commit f5f245cc, reversing
changes made to 94d37fdc.

this change broke gguf v2 which is incorrectly detected as big endian

Critical fix from llama.cpp JSON grammar to forbid un-escaped escape characters inside strings, which breaks parsing. (#3782)

* fix embedding by adding fixes from llama.cpp upstream

* remove assert

---------
Co-authored-by: Jesper Ek <deadbeef84@gmail.com>

This follows the same pattern for cuda and rocm to allow
disabling the build even when we detect the dependent libraries