Junyi Ma
Junyi Ma
Hello @smalltheater , Thanks for using our code! The setup of OverlapNet (pytorch) is the same as the one used in our newest work [OverlapTransformer](https://github.com/haomo-ai/OverlapTransformer) for faster retrieval, which is:...
Hi @ecr23xx and @chrischoy , have you ever created a function that can achieve this?
Thanks for your interest in our work! The code is released. We will gradually post more usage in Cam4DOcc.
We will release all the other tasks once everything is ready. Currently, please feel free to train and evaluate models for the most straightforward tasks (in the released version) related...
> Why not directly perform element-wise addition, which would also maintain the data dimensionality? Actually, we explicitly concatenate the features along the channel to increase the embedding dimension of the...
Thanks for your interest in our work! I am busy writing my master thesis but will release the code you need in one month. Your patience is appreciated :-)
Sorry for the late reply! You could refer to [pole-localization](https://github.com/PRBonn/pole-localization) to generate the poses at first. The release in this repo is coming.
Same question! Is there any plan to release ek55_eval_labels.pkl? Thanks!
Great! Thanks for sharing the annotation.
Hi @summmeer , Thanks for your great work. Have you tried to infer only with one or two steps rather than the total steps?