pretty_name: CDSM-Net Processed Seismic Dataset
license: cc-by-4.0
tags:
- seismic
- earthquake
- phase-picking
- event-detection
- waveform
- stead
- instance
- hdf5
- time-series
CDSM-Net Processed Seismic Dataset
This repository provides the processed waveform data, metadata splits, training summary, and evaluation results used in the CDSM-Net study for seismic event detection and P/S phase picking.
Files
STEAD/
merge.h5 # processed STEAD waveforms
train.csv # STEAD training split
val.csv # STEAD validation split
test.csv # STEAD test split
INSTANCE/
Instance.h5 # processed INSTANCE waveforms
test.csv # INSTANCE cross-domain test split
result.csv # final evaluation metrics
training_history.txt
README.md
The CSV files provide metadata and waveform indices used by the CDSM-Net code to read the corresponding HDF5 waveform files.
training_history.txt provides per-epoch validation metrics extracted from the
training log.
Data Sources
These files are processed derivatives of two public datasets:
- STEAD: STanford EArthquake Dataset
https://github.com/smousavi05/STEAD - INSTANCE: The Italian Seismic Dataset for Machine Learning
https://www.pi.ingv.it/banche-dati/instance/
This repository does not replace the original datasets. For complete raw data, station metadata, and official documentation, please refer to the original STEAD and INSTANCE releases.
License
This processed dataset is released under the Creative Commons Attribution 4.0 International License (CC BY 4.0), consistent with the licenses of the original STEAD and INSTANCE datasets.
Users should cite this processed dataset and the original STEAD and INSTANCE datasets when using these files.
Related Code
CDSM-Net source code and reproduction scripts are available at:
https://github.com/Fortunate-ziye/CDSM-Net
Citation
@misc{zhang2026cdsmnetdata,
title = {{CDSM-Net Processed Seismic Dataset}},
author = {Zhang, Hui and Wang, Shoubin and Peng, Guili and Zhang, Qi},
year = {2026},
publisher = {Hugging Face},
url = {https://huggingface.co/datasets/ziye0013/CDSM},
note = {Processed derivative dataset based on STEAD and INSTANCE}
}
Please also cite the original STEAD and INSTANCE dataset papers:
- Mousavi et al. (2019), IEEE Access, doi:
10.1109/ACCESS.2019.2947848 - Michelini et al. (2021), Earth System Science Data, doi:
10.5194/essd-13-5509-2021
Contact
Guili Peng
School of Control and Mechanical Engineering, Tianjin Chengjian University
Email: penguili@126.com