3 Commits

Author	SHA1	Message	Date
H1K0	632407ebef	refactor: split training scripts into pretrain.py and train.py ... - scripts/run_pretrain.py -> scripts/pretrain.py: pre-trains on McGill corpus (data/processed/mcgill/), saves checkpoints/pretrained.pt. - scripts/train.py: rewritten as high-level fine-tune wrapper; loads pretrained.pt, trains on data/processed/user/, saves finetuned.pt. Both scripts include timing estimate, loss-curve plot, per-epoch report, and --skip-training flag. - README: updated section 7 to reflect new script names and separate data directories. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-05-20 12:35:23 +03:00
H1K0	555205b7d2	docs: actualize vocab size (81→85), spec version (2.0→2.2), style tag (user→H1K0) ... Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-05-20 03:23:55 +03:00
H1K0	75fa07bf6c	docs: add README, architecture, glossary, requirements; update CLAUDE.md ... Add four Russian-language project documents: - README.md: user-facing guide (install, quick start, data prep, training, evaluation, limitations) - docs/architecture.md v1.0: system architecture, data flow diagrams, module interfaces, 7 architectural decision records, extension points - docs/glossary.md v1.0: musical, ML, and project-specific term definitions - docs/requirements.md v1.0: functional/non-functional requirements, acceptance criteria, four use-case scenarios Update CLAUDE.md with project name etymology (hamori / ハモリ) and rename repo root reference from chord-gen to hamori. Refine chord_format_spec.md. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-05-19 11:00:21 +03:00