AEmotionStudio/moss-music-models

MOSS-Music 8B — MAESTRO mirror

Verbatim mirror of OpenMOSS-Team/MOSS-Music-8B-Instruct (revision fce7f8304e96cc2d3398b8106456cbb2ecec3139, 2026-05) packaged for the MAESTRO DAW's offline runtime.

MOSS-Music 8B (OpenMOSS, Fudan University) is a music-understanding LLM — a Whisper-style audio encoder feeding a Qwen3-8B language model with DeepStack feature injection. It captions music, transcribes lyrics with timestamps, writes out chord progressions, maps song structure, names key and tempo, and answers free-form musical questions.

Layout

bf16/               everything MAESTRO downloads (manifest allow_patterns)
  model-0000{1..4}-of-00004.safetensors   bf16 shards, byte-identical to upstream
  model.safetensors.index.json
  config.json  generation_config.json
  tokenizer_config.json  vocab.json  merges.txt  added_tokens.json
  preprocessor_config.json  chat_template.jinja
  sha256.json       per-file hashes of this folder
README.md           this file

Differences from upstream

• The remote-code Python files (configuration_moss_music.py, modeling_moss_music.py, processing_moss_music.py) are not shipped: MAESTRO vendors the upstream GitHub src/ inference package (github.com/OpenMOSS/MOSS-Music @ ad107c7d) and never uses trust_remote_code.
• Weight shards are byte-identical to upstream (sha256-verified against the upstream LFS metadata at mirror time — see bf16/sha256.json).

License

Apache-2.0, same as upstream code and weights. Copyright the OpenMOSS team, Fudan University. This mirror exists so MAESTRO installs pull a pinned, verified snapshot; all credit for the model goes to its authors.