Difficulties with NumPy's np.argmax returning unexpected indices in multi-dimensional arrays
I'm converting an old project and I've spent hours debugging this and I've looked through the documentation and I'm still confused about I'm facing an issue with `np.argmax` when trying to find the index of the maximum value in a 3D NumPy array... I'm using NumPy version 1.23.5, and my goal is to retrieve the indices of the maximum values along the last axis of the array correctly. However, I'm getting unexpected results that don't seem to match the actual maximum values. Here’s the array I'm working with: ```python import numpy as np array = np.array([[[1, 2, 3], [4, 5, 6]], [[7, 8, 9], [10, 11, 12]], [[13, 14, 15], [16, 17, 18]]]) ``` When I run the following code: ```python max_indices = np.argmax(array, axis=-1) print(max_indices) ``` I expect to get an array that reflects the indices of the maximum values along the last axis for each 2D sub-array. Instead, I receive: ``` [[2 2] [2 2] [2 2]] ``` This output indicates the index `2` for each sub-array, which is confusing since the maximum values are actually in index `2`, but I thought it would return the indices of the max values relative to their sub-array, like `[2, 2]` for the last elements of each sub-array. I attempted to verify the results by checking the maximum values directly: ```python max_values = np.max(array, axis=-1) print(max_values) ``` This correctly gives me: ``` [ 3 6 18] ``` So, clearly, the maximum value indices are correct, but the index array is returning `2` for all. What am I missing here? Is there an issue with how I'm interpreting `np.argmax` in the context of multi-dimensional arrays? Any insights would be greatly appreciated! Thanks in advance! I'm using Python LTS in this project. I've been using Python for about a year now.