README.md

metadata

license: mit
configs:
  - config_name: gnps_random
    default: true
    data_files:
      - split: train
        path: GNPS/random/train.csv
      - split: validation
        path: GNPS/random/valid.csv
      - split: test
        path: GNPS/random/test.csv
  - config_name: gnps_scaffold
    data_files:
      - split: train
        path: GNPS/scaffold/train.csv
      - split: validation
        path: GNPS/scaffold/valid.csv
      - split: test
        path: GNPS/scaffold/test.csv
  - config_name: mist_canopus_split_0
    data_files:
      - split: train
        path: MIST_canopus/split_0/train_0.csv
      - split: validation
        path: MIST_canopus/split_0/val_0.csv
      - split: test
        path: MIST_canopus/split_0/test_0.csv
  - config_name: mist_canopus_split_1
    data_files:
      - split: train
        path: MIST_canopus/split_1/train_1.csv
      - split: validation
        path: MIST_canopus/split_1/val_1.csv
      - split: test
        path: MIST_canopus/split_1/test_1.csv
  - config_name: mist_canopus_split_2
    data_files:
      - split: train
        path: MIST_canopus/split_2/train_2.csv
      - split: validation
        path: MIST_canopus/split_2/val_2.csv
      - split: test
        path: MIST_canopus/split_2/test_2.csv
  - config_name: massbank_scaffold
    data_files:
      - split: train
        path: MassBank/scaffold/train.csv
      - split: validation
        path: MassBank/scaffold/valid.csv
      - split: test
        path: MassBank/scaffold/test.csv
  - config_name: massspecgym
    data_files:
      - split: train
        path: MassSpecGym/MassSpecGym_train.csv
      - split: validation
        path: MassSpecGym/MassSpecGym_valid.csv
      - split: test
        path: MassSpecGym/MassSpecGym_test.csv

NeurIPS 2026 mass spectrometry dataset

This repository was prepared automatically for Hugging Face Datasets.

Available configs

• gnps_random
• gnps_scaffold
• mist_canopus_split_0
• mist_canopus_split_1
• mist_canopus_split_2
• massbank_scaffold
• massspecgym

Usage

from datasets import load_dataset

ds = load_dataset("your-username/your-dataset-name", "gnps_random")

Replace the repo id and config name as needed.

Notes

• valid and val files are mapped to the Hugging Face validation split.
• .DS_Store files are ignored.