hamori

Files

T

H1K0 dd4f21f17f feat: add run_pretrain.py; fix output-path naming and max_seq_len

- scripts/run_pretrain.py: single-command pre-training runner with
  timing estimate, loss-curve plot (matplotlib), and per-epoch report.
  Sets max_seq_len=256 (McGill sequences max out at 195 tokens, ~4x
  faster attention than the 512 default).
- src/train.py: normalise --output so pretrained.pt and pretrained both
  produce pretrained.pt + pretrained.log.csv (not pretrained.pt.log.csv).
  Serialize Path fields as strings in checkpoint to satisfy weights_only.
- requirements.txt: drop unused pandas/music21, add mido (pretty_midi dep).

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

2026-05-20 12:13:38 +03:00

external_converters

feat: extend time signature support to 9 metres (5/4, 7/4, 7/8, 9/8)

2026-05-20 00:37:05 +03:00

__init__.py

chore: initialize project scaffold

2026-05-19 10:28:17 +03:00

chord_parser.py

feat: implement chord symbol parser with full test suite

2026-05-19 11:17:10 +03:00

dataset.py

feat: add dataset, prepare_data pipeline and fix McGill converter