Need help to resolve some issues when using sampling_time option in to_tensor() method in lava.lib.dl.slayer.io.Event() class

[gwgknudayanga]

Hello Experts,
*I have created as "td_event = Event(x_coordinate_vector,y_coordinated_vector,event_polarity_vector,timestamp_vector) using a 70ms length event stream from davis346 event camera(hence (346x260) spatial resolution and 2 polarities (intensity increment, intensity decrement).

• For that event stream first time stamp is 0 and last time stamp is 69363 (these are in microseconds).
• Since I want to divide this Event stream into 5 time bins, i set the sampling_time as "np.int(np.ceil(69363/5)) = 13873" of to_tensor method of Event() class executed the "td_tensor = td_event.to_tensor(sampling_time = 13873)". Then , But the resulting tensor "td_tensor" has the following shape.That is, the td_tensor has has 6 time bins instead 5 time bins which is wrong. So could you help to figure out the issue here?

---

td_tensor.shape = (2,260,346,6)

• Also, when the sampling_time = 1, the values which indicates the presence of event in above "td_tensor" is 1. But when the sampling_time = 13873, the values that indicate the presence of event is (1/12873) = , which doesn't have any meaning and if the sampling_time is increasing these values are closing to zero consequently making it difficult to indicate the presence of events in above "td_tensor". I feel this like a bug, Am i missing something? How can i avoid these problems?