Tune_gene about siamdw HOT 3 CLOSED

I rerun TPE many times and succeed finally, it is too slow. I then return to use GA with multi-GPUs by the following command:
mpiexec -n 8 python -u ./siamese_tracking/tune_gene.py
I notice that the speed of each thread(about 1 hour with OTB2013) is much slower than running GA with a single thread(20minutes).
Do you face the same problem when you test it? Any suggestions？

from siamdw.

I rerun TPE many times and succeed finally, it is too slow. I then return to use GA with multi-GPUs by the following command:
mpiexec -n 8 python -u ./siamese_tracking/tune_gene.py
I notice that the speed of each thread(about 1 hour with OTB2013) is much slower than running GA with a single thread(20minutes).
Do you face the same problem when you test it? Any suggestions？

I didn't meet that problem. I realized that you have opened several questions about tuning process. Email me for details may be better.

from siamdw.

I rerun TPE many times and succeed finally, it is too slow. I then return to use GA with multi-GPUs by the following command:
mpiexec -n 8 python -u ./siamese_tracking/tune_gene.py
I notice that the speed of each thread(about 1 hour with OTB2013) is much slower than running GA with a single thread(20minutes).
Do you face the same problem when you test it? Any suggestions？

I didn't meet that problem. I realized that you have opened several questions about tuning process. Email me for details may be better.

Thanks, I will send email to you if I have other questions about tuning.

from siamdw.