4 Commits

Include renames

Author	SHA1	Message	Date
silveroxides	3bd4977d2a	fix(cwb): keep preset scale neutral ... Runtime conditioning scale reduction must not be carried into persisted model or LoRA weights. Keep named merger presets at 1.0 and reserve scaling for explicit custom configuration.	2026-08-03 18:48:14 +02:00
silveroxides	89f4def83e	feat(cwb): preserve secondary-only layers ... - build generic and LoRA worklists from the union of inputs - copy single-source layers exactly and merge available providers - keep logical LoRA groups canonical, atomic, and collision-safe - retain mismatch, dtype, quantization, streaming, and OOM contracts	2026-08-02 20:26:41 +02:00
silveroxides	f390c12776	fix(merge): enforce LoRA tensor contracts ... - preserve direct diffs, precision, and guarded low-bit tensors - stream paired CWB layers with atomic CPU OOM fallback - save generated merge artifacts under models_dir	2026-08-02 18:11:19 +02:00
silveroxides	bed9c6e8ff	feat(merge): add dedicated CWB mergers	2026-08-01 13:14:14 +02:00