mirror of
https://github.com/ollama/ollama.git
synced 2026-11-22 00:01:45 +00:00
* mlxrunner: add Inkling 2 Mini * mlx: fuse Inkling 2 Mini residual mixing on Metal Fuse RMS normalization and query scoring in one kernel. Apply the softmax weights in a second kernel to avoid storing per-query products. * create: expose Inkling 2 Mini context length in model metadata Copy model_max_length to max_position_embeddings during import so ollama show reports the checkpoint's context length. Reject configs that specify conflicting limits. * mlxrunner: reduce Metal dispatches in Inkling 2 Mini decode Reuse the prefill mask shader to compute relative position bias in one dispatch for single-query attention. On M3 Ultra with Metal 3, compute the loop controller's gain and forget projections in one matrix multiply.
16 lines
427 B
Text
16 lines
427 B
Text
llama/**/*.cpp linguist-vendored
|
|
llama/**/*.hpp linguist-vendored
|
|
llama/**/*.h linguist-vendored
|
|
llama/**/*.c linguist-vendored
|
|
llama/**/*.cu linguist-vendored
|
|
llama/**/*.cuh linguist-vendored
|
|
llama/**/*.m linguist-vendored
|
|
llama/**/*.metal linguist-vendored
|
|
|
|
app/webview linguist-vendored
|
|
|
|
llama/build-info.cpp linguist-generated
|
|
|
|
* text=auto
|
|
*.go text eol=lf
|
|
model/renderers/testdata/inkling2_chat_template.jinja text eol=lf
|