Comments (8)
Please refer to here #23 for the issue about spconv 1.1
, where you could also find a spconv 1.1
version supported by @codyreading at https://github.com/codyreading/PCDet/tree/spconv-1.1.
from openpcdet.
Now PCDet
has supported spconv 1.1
as modified here 2244be4.
from openpcdet.
Hi, thank you for the amazing work. I tried installing spconv1., but their cuda libs search locations and their build process kept breaking on my lab's servers. I installed spconv1.1 instead, and it just worked. However, I noticed in several places this repo stresses spconv1.0. I am wondering if there is anything that absolutely needs spconv1.0.
The 2 main places this repo uses spconv is the
voxel_generator
and the convolutions.
I am stuck with the same issue,
I just cannot achieve a successful spconv v1.0 build on our lab's servers.
@monghimng: Were you finally able to build spconv v1.0 ?
See issue #70 for more details.
from openpcdet.
I just installed v1.1 and the branch the author provided. It somehow worked for me.
from openpcdet.
The repo has support for spconv1.1, the main difference is the GPU memory usage of the PartA2 ( a lot larger in spconv==1.1). Still should work fine from 2244be4.
from openpcdet.
I just installed v1.1 and the branch the author provided. It somehow worked for me.
@monghimng:
That's great news, if you are using conda, could you please provide the output of conda list
.
Then I could maybe just "clone" your conda environment and all is good.
@codyreading:
Thank you for pointing out the different GPU memory usage.
I tried both SECOND and PartA^2Net with batch_size=1
on a GeForce RTX 2080 Ti with 11GB and also on a TITAN Xp with 12GB GPU memory
and the training in both cases still crashed with unknown device type
as mentioned in issue #70.
Now, I think the issue most likely comes from my specific set of versions for
- python
- pytorch
- CUDA
- cudnn.
that I am using. For all details please see issue #70.
Sadly, I found no other combination yet that allowed both,
a successful build of spconv and a successful installation of PCDet.
from openpcdet.
Hi @MartinHahner88,
Doesn't look like I've seen that issue. I've been using the following versions which has worked for me, and using pip
with virtualenv
to install the packages.
- Python: 3.7
- Torch: 1.4
- CUDA: 10.1
- CUDNN: 7.6.5
from openpcdet.
Doesn't look like I've seen that issue. I've been using the following versions which has worked for me, and using
pip
withvirtualenv
to install the packages.* Python: 3.7 * Torch: 1.4 * CUDA: 10.1 * CUDNN: 7.6.5
Those versions, unfortunately, let to yet another issue for me, see here.
from openpcdet.
Related Issues (20)
- About BaseBEVBackbone HOT 2
- How to get validation loss and display it in Tensorboard HOT 4
- Abnormal demo.py result using nuscenses dataset pre-trained models HOT 3
- Ensemble or test-time augmentation? HOT 2
- free(): double free detected in tcache 2 HOT 2
- questions about GETTING_STARTED.md HOT 2
- 'EasyDict' object has no attribute 'BACKUP_DB_INFO' HOT 3
- What should we set the overlap values for custom datasets? HOT 5
- TypeError: expected str, bytes or os.PathLike object, not NoneType HOT 1
- `pillar_features` IndexError: too many indices for tensor of dimension 1 HOT 1
- Question about adam_onecycle optimizer HOT 2
- PV RCNN++ pretained model HOT 1
- [Docs] A100算力加持!书生大模型实战营第3期全面升级,趣味闯关模式等你开启 HOT 1
- How to change source of pypi?
- Projecting 3D bounding boxes on 2D image
- Floating Point Exception HOT 5
- Training/Testing pv-rcnn using custom data
- Custom Dataset npy file demo.py
- Error 700 illegal memory access encountered when training on custom data
- nuscenes_dataset process
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from openpcdet.