OpenPCDet

OpenPCDet copied to clipboard

wyddmw

Update augmentor_utils.py

1

Fix bug : "IndexError: too many indices for array: array is 1-dimensional, but 2 were indexed" Check that the data is not empty before operating the data

Number1JT

Update to reflect the config LR

2

Why is the LR hardcoded? Should it not be the same as that provided in the config file?

aastha3

Can not run the evaluation function successful

4

2022-09-22 09:36:49,204 INFO Total samples for Waymo dataset: 39987 2022-09-22 09:36:52,147 INFO ==> Loading parameters from checkpoint ../output/waymo_models/pv_rcnn_plusplus_resnet_d1/default/ckpt/checkpoint_epoch_30.pth to GPU 2022-09-22 09:36:52,486 INFO ==> Checkpoint trained from version: pcdet+0.5.2+0000000 2022-09-22...

sjtuljw520

The results provided by the evaluation function do not contain metrics in different distance. Can you improve the function which can provide results like follows: OBJECT_TYPE_TYPE_VEHICLE_LEVEL_1: [mAP 0.657237] [mAPH 0.652432]...

sjtuljw520

Adding new classes to a pre-trained model

4

Assuming the model was trained on kitti/waymo dataset with 3 classes, can we add new classes and continue training with the addition of these new classes? (by changing the dataset...

d33dler

The storage to store the processed waymo dataset?

7

The raw waymo dataset is about 1 T, and how much storage needed to store the processed waymo dataset?

sjtuljw520

Segmentation fault (core dumped) when run the demo.py

1

I run demo.py with own dataset (pv-rcnn pretrained model downloading from the codebase) and encouter this error: **Segmentation fault (core dumped)** After debuging, I located this error coming from: line...

sjtuljw520

@sshaoshuai Hello, I wonder what is the motivation for redefining a `DistributedSampler` [here](https://github.com/open-mmlab/OpenPCDet/blob/master/pcdet/datasets/__init__.py#L27), whose `__iter__` method is pretty similar to the PyTorch built-in DistributedSampler, as shown in the PyTorch source...

toyot-li

The results in paper are different to the Leaderboard of waymo dataset challenge website.

2

I find that the results in Table 3 of paper ("MPPNet: Multi-Frame Feature Intertwining with Proxy Points for 3D Temporal Object Detection") is different to the Leaderboard in waymo website....

sjtuljw520