Commit Graph

4 Commits

Author SHA1 Message Date
Patrick Devine
de5cb7311f mlx: add mxfp4/mxfp8/nvfp4 importing (#15015)
This change allows importing bf16 and converting to mxfp4/mxfp8/nvfp4
and also importing fp8 and converting directly to mxfp8.
2026-03-24 13:45:44 -07:00
Daniel Hiltgen
10e51c5177 MLX: add header vendoring and remove go build tag (#14642)
* prefer rocm v6 on windows

Avoid building with v7 - more changes are needed

* MLX: add header vendoring and remove go build tag

This switches to using a vendoring approach for the mlx-c headers so that Go
can build without requiring a cmake first.  This enables building the new MLX
based code by default.  Every time cmake runs, the headers are refreshed, so we
can easily keep them in sync when we bump mlx versions.  Basic Windows
and Linux support are verified.

* ci: harden for flaky choco repo servers

CI sometimes fails due to choco not actually installing cache.  Since it just speeds up the build, we can proceed without.

* review comments
2026-03-09 17:24:45 -07:00
Jesse Gross
f20dc6b698 mlx: don't default to affine quantization for unquantized models
Otherwise the BF16 version of models trigger segfaults when they
call into quantized kernels.
2026-02-23 15:03:53 -08:00
Patrick Devine
041fb77639 model: add gemma3 to the mlxrunner (#14276)
This change adds the gemma3 model to the mlxrunner and simplifies some of the quantization
code for loading weights.
2026-02-15 22:47:59 -08:00