I don't understand how to translate them to the initial state . Tks

This is 4 layered Rubics Cube. These layers are situated behind each other (perpendicularly). You can think them as 4 Cards placed behind each other.
Also, each layer is transparent. So, color on the second, third or fourth layer can be seen through the first layer.

Top View is watching the cube from the first layer. As all the layers are transparent, so we can see the colors of each layer behind it from the first layer.

So, from the TOP VIEW you will see a complete cube combining each layer's colors.

In the description, the first picture is the 4 layers of the cube. Can you place them behind each other? What will you see if you observe them from the first layer and considering them as transparent?
If you see them like below, then bingo you get the idea!!

YRYB
RRYY
GGBB
RGBG

Here, the first Y and the last B color of the first row are coming from the 4th layer if you see. And so on for the rests.

In the given input, let say the color code is like below,
1 - Red
2 - Yellow
3 - Blue
4 - Green.

So, the input basically says that what you will see all the layers from the front of the first layer.
If you replaced the number with the color, then you will get the exact same cube that is given in the description. (We have already built it above!!)