First of all, thank you for sharing this excellent work — both the code and the accompanying blog are insightful and inspiring. I greatly appreciate the clarity and depth of the explanations.
I'm currently exploring your implementation and found the visualizations in the blog post extremely helpful for understanding the results and intuitions behind NSA. To better reproduce and study the experiments, would you consider sharing the code (or Jupyter notebooks) used to generate those visualizations?
Having access to those would significantly improve the reproducibility and learning experience for others who are studying the method.
Thanks again for your outstanding contribution to the community!
First of all, thank you for sharing this excellent work — both the code and the accompanying blog are insightful and inspiring. I greatly appreciate the clarity and depth of the explanations.
I'm currently exploring your implementation and found the visualizations in the blog post extremely helpful for understanding the results and intuitions behind NSA. To better reproduce and study the experiments, would you consider sharing the code (or Jupyter notebooks) used to generate those visualizations?
Having access to those would significantly improve the reproducibility and learning experience for others who are studying the method.
Thanks again for your outstanding contribution to the community!