superkoma
superkoma
Hi, I've tried to re-train DAMSM on coco dataset. But the s_loss has decreased to 0.53 & 0.56, w_loss has decreased to 0.35 & 0.40 since 60 epochs without any...
Hi, Thanks for the interest for our work. We are processing the codes of preprocessing dataset now and will upload it as soon as possible.
Hi, The data used in our experiments is uploaded, and some codes of preprocessing are under cleaning.
Please check the updated information in the main content of README.
Have you finished the ground-truth sampling data creation before?
I have encountered this issue before, could you please provide the version of your installed `waymo-open-dataset-tf`, such as `2-6-0` or `2-0-0`?
- After changing the code structure to OpenPCDet, we have been preparing the model under the new structure and will release it according to Waymo's requirements. - I am still...
It's used to align the format to assign the corresponding supervised information. 200 is suitable for WOD, it can be changed to assign your own data setting.
Could you provide the details of your question?
- After changing the code structure to OpenPCDet, we have been preparing the model under the new structure and will release it according to Waymo's requirements. - I am still...