smousavi05
commented
1 year ago this is more likely due to your selected overlappibg window and batch size. You can either play with the batch size or pad all traces with a few hour-long zeros.
On Wed, Oct 5, 2022, 6:34 PM Alireza Niksejel @.***> wrote:
Dear @smousavi05 https://github.com/smousavi05
Recently, I downloaded 1 year of continuous data from IRIS and used EQT for phase identification. Everything goes well regarding the picking process however, I noticed that there are no picks detected after 20:00 on all waveforms. To further explore the issue, I selected a set of 3 component 1-day-long data from one station and cut it from 20:00:00 to 23:59:59 and surprisingly EQT generated many picks out of it. I've run the EQT directly on mseed data and also HDF5 files and got the same issue. Do you have any suggestions/ideas on where this could be generated from?
— Reply to this email directly, view it on GitHub https://github.com/smousavi05/EQTransformer/issues/144, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADHWAPGRJJZEUQ5PHVXKZLDWBWNZXANCNFSM6AAAAAAQ5WBQSY . You are receiving this because you were mentioned.Message ID: @.***>
Dear @smousavi05
Recently, I downloaded 1 year of continuous data from IRIS and used EQT for phase identification. Everything goes well regarding the picking process however, I noticed that there are no picks detected after 20:00 on all waveforms. To further explore the issue, I selected a set of 3 component 1-day-long data from one station and cut it from 20:00:00 to 23:59:59 and surprisingly EQT generated many picks out of it. I've run the EQT directly on mseed data and also HDF5 files and got the same issue. Do you have any suggestions/ideas on where this could be generated from?