organization of datasets

[scarLiu]

It is commendable that your work has been very meticulous, but if you can list the organization format of each file in your Kitti dataset, it will be more friendly to beginners~

[scarLiu]

and how can I use the pertained model in the process of evolution? 😊

[SenZHANG-GitHub]

It is commendable that your work has been very meticulous, but if you can list the organization format of each file in your Kitti dataset, it will be more friendly to beginners~

I use the original format of KITTI except for the matched IMU files. The details can be found in https://journals.sagepub.com/doi/pdf/10.1177/0278364913491297

[SenZHANG-GitHub]

and how can I use the pertained model in the process of evolution? blush

We can use the argument --load_weights_folder to load the pretrained model : )

[scarLiu]

and how can I use the pertained model in the process of evolution? blush

We can use the argument --load_weights_folder to load the pretrained model : )

Thank you for reply! And now I am troubled by a problem, that is, when the imu between two img frames is missing, how do you process the aligned data and get the predicted poses?

[SenZHANG-GitHub]

and how can I use the pertained model in the process of evolution? blush

We can use the argument --load_weights_folder to load the pretrained model : )

Thank you for reply! And now I am troubled by a problem, that is, when the imu between two img frames is missing, how do you process the aligned data and get the predicted poses?

Hi @scarLiu , I simply mark it as 'nan' in the matched imu files, and ignore those pairs with missing imu during training. Please note that we do not require imu during inference : )