search

DerrickXuNu / OpenCOOD

[ICRA 2022] An opensource framework for cooperative detection. Official implementation for OPV2V.
https://mobility-lab.seas.ucla.edu/opv2v/
Other
663 stars 99 forks source link

How to adjust the communication Volume #98

Closed HuangZhe885 closed 1 year ago

HuangZhe885 commented 1 year ago

Thank you for your excellent work. I use the formula log2 ( |M(k)| × D × 32/8) to calculate communication Volume, the calculation result is about 23~25. What should I do if I wish to manually lower this value? minimize the compression ratio or the feature size? I tried changing the compression ratio, but the Volume didn't go down.

image
DerrickXuNu commented 1 year ago

Sorry for the late response. Can you paste your code of calculating volume here?

HuangZhe885 commented 1 year ago

Sorry for the late response. Can you paste your code of calculating volume here? Thank you for your reply, I wrote this using the formula in the where2comm paper, but I'm not sure if it's correct.

non_zero_elements = communication_masks.nonzero().size(0) communication_volume = math.log2(non_zero_elements * C * 32 / 8) print("communication_volume",communication_volume)

DerrickXuNu commented 1 year ago

Changing compression rate should work as it will reduce the C. Please check if you are giving the right C (i.e., the feature dimension after compression ) to your formula.

besma14 commented 5 months ago

I am unable to reduce the communication volume even when adjusting the compression rate in where2comm. Could you help me?