search

Haiyang-W / DSVT

[CVPR2023] Official Implementation of "DSVT: Dynamic Sparse Voxel Transformer with Rotated Sets"
https://arxiv.org/abs/2301.06051
Apache License 2.0
373 stars 28 forks source link

DSVT all in tensorRT #52

Closed jingyue202205 closed 1 year ago

jingyue202205 commented 1 year ago

I implemented DSVT using the Tensorrt API and plugin！ If anyone is interested, the repository is (https://github.com/jingyue202205/DSVT-AI-TRT)

Haiyang-W commented 1 year ago

I implemented DSVT using the Tensorrt API and plugin！ If anyone is interested, the repository is (https://github.com/jingyue202205/DSVT-AI-TRT)

Your tremendously valuable efforts are greatly appreciated. In all honesty, deployment isn't our forte. Your assistance, however, will undoubtedly simplify the deployment process for DSVT, making the onboarding of the Transformer much smoother.

We also extend an invitation for you to stay tuned to our latest work UniTR (https://github.com/Haiyang-W/UniTR). This work extends DSVT into the domain of multi-modal 3D perception and introduces weight-sharing among multiple modalities (e.g., LiDAR and Cameras). It's the First Unified Multi-modal Transformer Backbone for 3D Perception. We're actively working on it and aiming to release the code before ICCV.

DSVT and UniTR might serve as preliminary steps towards the development of a foundation model for autonomous driving perception. With your invaluable support, I am confident we are moving in the right direction.

Thanks for your efforts once again :).

Haiyang,