ml-explore/mlx

[Feature Request] Fallback for custom kernels

クローズ

#3,240 opened on 2026/03/11

 (4 件のコメント) (0 件のリアクション) (0 人の担当者)C++ (1,794 件のフォーク)batch import
enhancementgood first issue

Repository metrics

Stars
 (26,216 個のスター)
PR merge metrics
 (平均マージ 6d 17h) (30d で 75 merged PRs)

説明

We have a specific use-case whereby we need to export MLX callables from Linux. Currently, this prevents us from using any operations that require custom kernels (e.g. grid sample) because otherwise, MLX complains that no Metal GPU was found.

Ideally, we would want to keep the fast kernel so that we can export without a Metal GPU, but execute on a Metal GPU.

コントリビューターガイド