ChiWeiHsiao

Hi, you can download trained weight [here](https://drive.google.com/file/d/1z_QbT8vU4XLX5Jx8SC7fxbovuGD2Dlqs/view?usp=sharing).

Hello, Sorry I don’t have access to my computer now, I will upload the weight on 19th :)

Hi @shahabty @Yusufma03, The trained weights can be downloaded [here](https://drive.google.com/drive/folders/1Zb6wObjdZ2lvhM07pgvJGhoZgJKFJMOa?usp=sharing). This is trained for about 400 epochs, but it also suffer from overfitting. :( The author of DeepVO paper seems...

Hi @yuanjinsheng, The files are large, maybe the remaining space of your computer is not enough? Could you post your error message?

@yuanjinsheng You can download all image datas and ground truth pose [here](http://www.cvlibs.net/datasets/kitti/eval_odometry.php), email register is needed. And the urls in script are from [here](http://www.cvlibs.net/datasets/kitti/raw_data.php), choose [synced+rectified data].

Hi, you can download the trained model here: https://drive.google.com/file/d/1l0s3rYWgN8bL0Fyofee8IhN-0knxJF22/view

Hi @ujgygy656565 , Please see [this](https://github.com/ChiWeiHsiao/DeepVO-pytorch/issues/6#issuecomment-473794631) and [this response](https://github.com/ChiWeiHsiao/DeepVO-pytorch/issues/6#issuecomment-473801815).

Hi @dasing, Since that position is the center of the kernel, it should be filled with the same pixel as a conventional kernel. (in line 57 ```new_result[1, 1] = (img_r,...

Hi @xuwei1245645564, 1) Yes, you can use larger image size, I reduce the image size just because memory issue. 2) Sorry I cannot recall precisely why I choose different sequences...

I am not sure whether using ratio 1:2 will result to better result than 1:1. You could try the 1:2 setting by specifying the outshape argument in line 119, 120...