Skip to main content

Ctrl+K

Increase citations, ease review & foster collaboration

Motivation

Why make it reproducible?
Increase Citations
Ease Review
Foster Collaboration

Front Matter

Installation
- Using Conda
- Conda on Apple M1 Chip
Data
- Data Preparation Notebook

How to make ML Reproducible

1. Model Evaluation
- 1.3. Evaluation Notebook
- 1.4. Evaluation Resources
2. Benchmarking
- 2.3. Benchmarking Notebook
- 2.4. Benchmarking Resources
3. Model Sharing
- 3.3. Sharing Notebook
- 3.4. Sharing Resources
4. Testing
- 4.4. Testing Notebook
- 4.5. Testing Resources
5. Interpretability
- 5.3. Interpretability Notebook
- 5.4. Interpretability Resources
6. Ablation Studies
- 6.2. Ablation Notebook
- 6.3. Ablation Resources

Talks & Workshops

Increase citations, ease review & collaboration – Making machine learning in research reproducible
Real-world Perspectives to Avoid the Worst Mistakes using Machine Learning in Science

Backmatter

References
Cite this Book

.md

Model Sharing Resources

Contents

3.4.1. Bibliography

3.4. Model Sharing Resources#

Currently sources of randomness are documented in Pytorch and Keras.

The resource The Turing Way further goes into detail on code quality and good data practices for reproducible research [Community et al., 2019].

3.4.1. Bibliography#

The Turing Way Community, Becky Arnold, Louise Bowler, Sarah Gibson, Patricia Herterich, Rosie Higman, Anna Krystalli, Alexander Morley, Martin O'Reilly, and Kirstie Whitaker. The turing way: a handbook for reproducible data science. 2019. URL: https://zenodo.org/record/3233986, doi:10.5281/ZENODO.3233986.

previous

3.3. Model Sharing

next

4. Testing

Contents

3.4.1. Bibliography

By Jesper Dramsch

© Copyright 2022 – 2024.

Terms & Privacy Policy | Maintained by Jesper Dramsch, creator of PythonDeadlin.es, Data-Science-Gui.de, Late to the Party 🎉 and the Latent Space Community.