LiT: Unifying LiDAR "Languages" with LiDAR Translator
Authors: Yixing Lao, Tao Tang, Xiaoyang Wu, Peng Chen, Kaicheng Yu, Hengshuang Zhao
NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our experiments are conducted on three widely utilized datasets for autonomous driving research: KITTI [9], Waymo [10], and nu Scenes [11]. These datasets present a diverse range of Li DAR characteristics, including varying beam counts and sensor configurations, making them ideal for evaluating our Li DAR translation approach, Li T. |
| Researcher Affiliation | Collaboration | Yixing Lao The University of Hong Kong yxlao@cs.hku.hk Tao Tang Sun Yat-sen University ttang@mail.sysu.edu.cn Xiaoyang Wu The University of Hong Kong xywu@cs.hku.hk Peng Chen Cainiao Group yuanshang.cp@cainiao.com Kaicheng Yu Westlake University yukaicheng@westlake.edu.cn Hengshuang Zhao The University of Hong Kong hszhao@cs.hku.hk |
| Pseudocode | No | The paper describes the steps and modules of the Li DAR Translator but does not include any formal pseudocode or algorithm blocks. |
| Open Source Code | Yes | Source code and demos are available at: https://yxlao.github.io/lit. |
| Open Datasets | Yes | Our experiments are conducted on three widely utilized datasets for autonomous driving research: KITTI [9], Waymo [10], and nu Scenes [11]. |
| Dataset Splits | Yes | Our experiments are conducted on three widely utilized datasets for autonomous driving research: KITTI [9], Waymo [10], and nu Scenes [11]. These datasets present a diverse range of Li DAR characteristics... The evaluations are conducted using the SECOND-Io U [50] and PV-RCNN [62] models, following the evaluation protocol of previous works [36, 7, 51]. |
| Hardware Specification | Yes | The efficiency of this engine allows Li T to translate a multi-frame Li DAR scene typically in less than one minute, making large-scale domain unification tasks feasible. For more detailed runtime statistics, refer to Table 6. ... Runtime is measured on a single NVIDIA RTX 4090 GPU. |
| Software Dependencies | No | The paper mentions "Intel Embree" and "Nvidia Opti X" as technologies used for acceleration but does not provide specific version numbers for these or other software dependencies like programming languages or libraries (e.g., Python, PyTorch versions). |
| Experiment Setup | Yes | The hyperparameters for scene modeling are shown in Table 7. For background modeling, while most parameters are consistent across both Waymo and nu Scenes datasets, a key difference is their frame rates. ... We provide hyperparameters used for training in Table 8. |