go-mlx/internal
Snider fb0692baf3 test(metal): add LoRA end-to-end training pipeline test (Phase 3)
Validates full pipeline: load Gemma3-1B → apply LoRA (rank=8, 745K
params across 52 layers) → train 5 steps with cross-entropy loss
(7.15→6.31) → save adapter to safetensors → reload and verify all
weights match. Uses ValueAndGrad for autograd + AdamW optimiser.

Co-Authored-By: Virgil <virgil@lethean.io>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-19 23:09:16 +00:00
..
metal test(metal): add LoRA end-to-end training pipeline test (Phase 3) 2026-02-19 23:09:16 +00:00