When I'm do not generate in a for logp, what does the generated event represents?

[DachunKai]

Hi, I just want to generate in this manner as below:

events = esim.forward(
    log_images,        # torch tensor with type float32, shape T x H x W
    timestamps_ns  # torch tensor with type int64,   shape T 
)

and I get one event. Does this describe events between $ I_0 $ and $ I_{T-1} $? How do you merge the sub_event that generate in a for loop?