SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2407.06084 · IJCAI · 2024

3D Vision and Language Pretraining with Large-Scale Synthetic Data

Yang Liu, Siyuan Huang, Qingchao Chen, Wentao Mo, Dejie Yang, Zhu Xu

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 10 functions out of this paper's own repositories and ran 10 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
idejie/3DSyn canonical 10 of 10
FunctionStatusWhere it lives
in_hull Ran idejie/3DSyn/scannet/model_util_scannet.py
code served (permissive licence) · get_code("aab935d4b7b55c2c")
compute_normal Ran idejie/3DSyn/scannet/scannet_utils.py
code served (permissive licence) · get_code("be1fe3da19cb1fd1")
extract_pc_in_box3d Ran idejie/3DSyn/scannet/model_util_scannet.py
code served (permissive licence) · get_code("50bdbf961c28b4b2")
normalize_v3 Ran idejie/3DSyn/scannet/scannet_utils.py
code served (permissive licence) · get_code("e718b6389b360792")
random_caption_word Ran idejie/3DSyn/dataset/data_converter.py
code served (permissive licence) · get_code("f68277b7b645f232")
random_point_cloud Ran idejie/3DSyn/dataset/data_converter.py
code served (permissive licence) · get_code("1e399fcc55e7f292")
read_aggregation Ran idejie/3DSyn/scannet/load_scannet_data.py
code served (permissive licence) · get_code("7f50e9ea5a5a0a0a")
read_segmentation Ran idejie/3DSyn/scannet/load_scannet_data.py
code served (permissive licence) · get_code("f124539052ae9169")
represents_int Ran idejie/3DSyn/scannet/scannet_utils.py
code served (permissive licence) · get_code("2857904cc41e5290")
rotate_aligned_boxes Ran idejie/3DSyn/scannet/model_util_scannet.py
code served (permissive licence) · get_code("7457a44fbe4c2d1d")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

3D Vision-Language Pre-training (3D-VLP) aims to provide a pre-train model which can bridge 3D scenes with natural language, which is an important technique for embodied intelligence. However, current 3D-VLP datasets are hindered by limited scene-level diversity and insufficient finegrained annotations (only 1.2K scenes and 280K textual annotations in ScanScribe), primarily due to the labor-intensive of collecting and annotating 3D scenes. To overcome these obstacles, we construct SynVL3D, a comprehensive synthetic scenetext corpus with 10K indoor scenes and 1M descriptions at object, view, and room levels, which has the advantages of diverse scene data, rich textual descriptions, multi-grained 3D-text associations, and low collection cost. Utilizing the rich annotations in SynVL3D, we pre-train a simple and unified Transformer for aligning 3D and language with multi-grained pretraining tasks. Moreover, we propose a synthetic-to-real domain adaptation in downstream task fine-tuning process to address the domain shift. Through extensive experiments, we verify the effectiveness of our model design by achieving state-of-the-art performance on downstream tasks including visual grounding, dense captioning, and question answering. Codes are available at: https://github.com/idejie/3DSyn.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2407.06084")
get_code_for_paper("2407.06084")
have("2407.06084")

Connect an agent — have() is free.