about training data of in-context editing in Figure 6

[garychan22]

Hi, thanks for the great work here.

which subset in https://huggingface.co/datasets/OmniGen2/X2I2 corresponds to training data of in-context editing in Figure 6? I have checked the jsonls but it seems that there are no multiple input_images for this task.

thanks!

[CS123n]

Same question for the multi-object in-context generation dataset