Native Apple Metal GPU inference via mlx-c bindings

Snider a44e9f5789 feat(metal): add inference metrics (timing, throughput, memory) ... Instrument Generate, Classify, and BatchGenerate with: - Prefill/decode timing (separate phases) - Token counts (prompt + generated) - Throughput (tok/s for each phase) - Peak and active GPU memory via Metal allocator Wire through metalAdapter.Metrics() to go-inference interface. Test validates all fields populated after generation. Gemma3-1B 4-bit on M3 Ultra: prefill 246 tok/s, decode 82 tok/s, peak 6.2 GB GPU memory. Co-Authored-By: Virgil <virgil@lethean.io> Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>		2026-02-19 23:34:40 +00:00
cpp	fix(metal): address 4 minor code review items	2026-02-19 21:36:40 +00:00
docs/plans	docs: batch inference API design (Phase 5)	2026-02-19 23:18:38 +00:00
internal/metal	feat(metal): add inference metrics (timing, throughput, memory)	2026-02-19 23:34:40 +00:00
.gitignore	chore: gitignore dist/ (CMake install output)	2026-02-19 19:30:23 +00:00
CLAUDE.md	feat(api): migrate to go-inference shared interfaces	2026-02-19 20:15:42 +00:00
CMakeLists.txt	feat: extract go-mlx from go-ai as standalone Metal inference package	2026-02-19 17:57:37 +00:00
FINDINGS.md	fix(metal): address 4 minor code review items	2026-02-19 21:36:40 +00:00
go.mod	feat(api): migrate to go-inference shared interfaces	2026-02-19 20:15:42 +00:00
mlx.go	feat(api): migrate to go-inference shared interfaces	2026-02-19 20:15:42 +00:00
mlx_stub.go	feat: extract go-mlx from go-ai as standalone Metal inference package	2026-02-19 17:57:37 +00:00
mlx_test.go	feat(metal): add inference metrics (timing, throughput, memory)	2026-02-19 23:34:40 +00:00
register_metal.go	feat(metal): add inference metrics (timing, throughput, memory)	2026-02-19 23:34:40 +00:00
TODO.md	docs: mark batch inference API complete in TODO.md	2026-02-19 23:28:40 +00:00