Camera configuration in dataset_img_keys.json must be 4 caemra?

[Wonder1905]

Please help me understand this format:
"rh20t": {
"image_keys": [
"image",
"image",
"image",
"image"
],
"image_mask":[1,0,0,0]
},
"calvin": {
"image_keys": [
"rgb_static",
"rgb_gripper",
"rgb_gripper",
"rgb_gripper"
],
"image_mask":[1,1,0,0]
},
"bridgev2": {
"image_keys": [
"images0",
"images0",
"images0",
"images0"
],
"image_mask":[1,0,0,0]
}
From "configs/dataset_img_keys.json"
Lets say I have 2 image, left and right cameras, should I still keep the four images format?
Should I mask the rest?
i.e the model always need to get 4 images, if I have only two, It will add background images?
Thanks!