I'd like to ask you about implementing your code

[soohyoen]

Thank you for your excellent work.

I read your paper and found out that the amazing method improves the performance even in real-world data.

Therefore, I would like to analyze your code entirely.

When analyzing your paper, it was confirmed that the absolute value of the gradient was added to split in the direction of the largest gradient when it was higher than threshold.

May I know which part of your code has been modified?

[TY424]

Thank you for your excellent work.

I read your paper and found out that the amazing method improves the performance even in real-world data.

Therefore, I would like to analyze your code entirely.

When analyzing your paper, it was confirmed that the absolute value of the gradient was added to split in the direction of the largest gradient when it was higher than threshold.

May I know which part of your code has been modified?

Thank you, the core modification in the code is here: Homodirectional Gradient.Furthermore, it is necessary to update the parameter list of the associated function. Incorporating reduce_opacity() is also recommended to further enhance performance.