Open Amelia2675 opened 2 months ago
Hi Amelia, Thanks for checking out our repo. Could you please use np.nonzero to check if it's really an all zero array? It could be that the beginning and end are mostly zero which is very normal for event based data.
Dear ZuoWen,
Thank you for your response.
I have noticed that the data shape
and extended_data shape
appear as expected, but it seems that there is no data present in the extended_data
.
As stated in the README
, data_dir
is directed to "/DATA/pupil_st/data_ts_500"
. I am uncertain about what exactly should be placed in the data_ts_500
directory. Could you please provide further guidance on this?
Thank you!
Sincerely, Amelia
hi Amelia, for using the convlstm-et-pytorch-event.py script, the data directories are organized as: data_dir_train = "./dataset/data_ts_pro/train/" data_dir_val = "./dataset/data_ts_pro/val/" target_dir = "./dataset/pupil_st/label"
The data are preprocessed event frames which can be downloaded from (link available in the readme): https://drive.google.com/drive/folders/16qH_wv_oVNysJARtHIUrIXbHjOygfq_i?usp=sharing
The process_event.py script is to process raw events into event frame, and for that you need to download data from: https://drive.google.com/drive/folders/1HeOS5YBLruzHjwMKyBQfVTc_mJbsy_R1
Please let me know if you have any further question! best, zw
Hello,
Thank you for your contribution!
When I executed the
process_event.py
, I encountered some issues, as shown in the picture.I have two questions:
process_event.py
script, why doesh5_file.root.vector[:]
not capture the data in the h5_file? (It seems that the array contains all zeros)Thank you.
Best regards, Amelia