Instructions to use m-a-p/SheetSage2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use m-a-p/SheetSage2 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("feature-extraction", model="m-a-p/SheetSage2", trust_remote_code=True)# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("m-a-p/SheetSage2", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
License for the Python code in SheetSage2 and MERT-v2 repositories
Hi, and thank you for releasing SheetSage2 and MERT-v2.
I maintain YuE2-ComfyUI, an Apache-2.0 node pack for ComfyUI built around YuE2. I would like to add audio-to-score transcription with SheetSage2, so users can cover a recording without a separate Python environment.
I could not find which license covers the code in these repositories:
m-a-p/SheetSage2: the model and notation modules, infer.py
m-a-p/MERT-v2-FullSong: modeling_mert2.py, configuration_mert2.py
The LICENSE files cover the model weights (CC BY-NC 4.0) and say the code is licensed separately. THIRD_PARTY_NOTICES.md lists the dependencies, but neither file names a license for the repositories' own Python code. The YuE repository itself is Apache-2.0, and the pack already includes its yue2 sources unmodified under that license.
Could you tell me:
Is the Python code in these two repositories under Apache-2.0, like YuE, or under another license?
If it is Apache-2.0, may I include those files unmodified in the pack, with your copyright notice and a link back? The weights would stay a separate download under CC BY-NC 4.0, and the pack's README already states that non-commercial term for YuE2.
If the answer is yes, a short license note in the model cards would help anyone else doing the same.
Thanks!