MS-GPT: Rethinking MS/MS De Novo Structure Elucidation as Spectrum-Induced Posterior Querying of a Molecule-Language Model
Paper • 2607.23607 • Published
How to use nielsr/MS-GPT-SAFE-GPT-Base with Transformers:
# Load model directly
from transformers import AutoTokenizer, AutoModel
tokenizer = AutoTokenizer.from_pretrained("nielsr/MS-GPT-SAFE-GPT-Base")
model = AutoModel.from_pretrained("nielsr/MS-GPT-SAFE-GPT-Base", device_map="auto")This repository contains the SAFE-GPT base checkpoint included in the released asset bundle for MS-GPT: Rethinking MS/MS De Novo Structure Elucidation as Spectrum-Induced Posterior Querying of a Molecule-Language Model.
The included configuration identifies the original base as
datamol-io/safe-gpt.
These files are mirrored from the authors' released MS-GPT asset bundle with their explicit authorization.
This checkpoint retains the upstream Apache-2.0 license, as confirmed by the MS-GPT authors in GitHub issue #1.