Not All Points Are Equal: Learning Highly Efficient Point-based Detectors for 3D LiDAR Point Clouds (CVPR 2022, Oral)
Apache License 2.0
379
stars
50
forks
source link
How many memory do you need for waymo, I downloaded the data set and find it huge as 800g, btw, how long did you train with waymo and how many 2080tis you used #11
From my experience, the Waymo dataset is more than 1 TB including 800 GB training set and 200GB validation set. And you need at least another block of storage of the same size to prepare the dataset for OpenPCDet.
We train our model on 8 RTX 2080Ti GPUs and it takes around 12 hours on 20% data (~32k frames) of all the training samples.
yanghu819
commented
2 years ago
From my experience, the Waymo dataset is more than 1 TB including 800 GB training set and 200GB validation set. And you need at least another block of storage of the same size to prepare the dataset for OpenPCDet.
We train our model on 8 RTX 2080Ti GPUs and it takes around 12 hours on 20% data (~32k frames) of all the training samples.