Colormaps are not matching with reference pngs

Nobody has claimed this yet.

### Bug report

I've recently found out that built-in colormaps doesn't exactly match the ones shown on the [colormap reference](https://matplotlib.org/stable/gallery/color/colormap_reference.html).

Here's an example for the viridis colormap :

Last color on the png is `#fde724`:

Last color of the colormap is `#fde725`:
```python
In [35]: import matplotlib
...: from pylab import cm
...: colormap = cm.get_cmap("viridis")
...: matplotlib.colors.rgb2hex(colormap(colormap.N-1))
Out[35]: '#fde725'
```
I've tested only one other colormap (jet), and it had the same issue. Is there a reason for it ?

**Matplotlib version**

* Operating system: Ubuntu 20.04
* Matplotlib version 3.3.0 & 3.4.2
* Matplotlib backend TkAgg
* Python version: 3.8

Start by reproducing the reported discrepancy with Python 3.8, Matplotlib 3.3.0 or 3.4.2, and the `cm.get_cmap("viridis")` entry point. Compare the generated reference image with `matplotlib.colors.rgb2hex(colormap(colormap.N-1))`; done means identifying and documenting why the displayed and built-in colors differ, with a regression check if an appropriate test location is found.

Written by the indexing model from the issue text.