SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2401.10226 · CVPR · 2024

Towards Language-Driven Video Inpainting via Multimodal Large Language Models

Shangchen Zhou, Ziwei Liu, Yunhai Tong, Kai Chen, Chen Loy, Xiangtai Li, Jingkang Yang, Yining Li, Jiangning Zhang, Jianzong Wu, Chenyang Si

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 10 functions out of this paper's own repositories and ran 7 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
jianzongwu/Language-Driven-Video-Inpainting canonical 7 of 10
FunctionStatusWhere it lives
create_1d_absolute_sin_cos_embedding Ran jianzongwu/Language-Driven-Video-Inpainting/rovi/models/utils.py
pointer only (licence: NONE) · get_code("2eb56328ee222a96")
get_cosine_scheduler Ran jianzongwu/Language-Driven-Video-Inpainting/rovi/util.py
pointer only (licence: NONE) · get_code("d38dd5fa6084deeb")
get_mask_loss_weight_scale_quadratic Ran jianzongwu/Language-Driven-Video-Inpainting/rovi/util.py
pointer only (licence: NONE) · get_code("a37c74cff74be0d0")
get_unet_constant_deocder_cosine_scheduler Ran jianzongwu/Language-Driven-Video-Inpainting/rovi/util.py
pointer only (licence: NONE) · get_code("5ca301a24eda4247")
inflated_interpolate_3d Ran jianzongwu/Language-Driven-Video-Inpainting/rovi/models/mask_decoder.py
pointer only (licence: NONE) · get_code("1b457a7bd99a3de9")
load_video Ran jianzongwu/Language-Driven-Video-Inpainting/inference_interactive.py
pointer only (licence: NONE) · get_code("a612419f9fdb01ba")
transform_first_frame Ran jianzongwu/Language-Driven-Video-Inpainting/inference_interactive.py
pointer only (licence: NONE) · get_code("0b8d172f39477909")
get_down_block Not yet run jianzongwu/Language-Driven-Video-Inpainting/rovi/models/unet_blocks.py
pointer only (licence: NONE) · get_code("f6cbe1f5a30557f3")
get_mask_decoder Not yet run jianzongwu/Language-Driven-Video-Inpainting/rovi/models/mask_decoder.py
pointer only (licence: NONE) · get_code("45965320324d4cff")
get_up_block Not yet run jianzongwu/Language-Driven-Video-Inpainting/rovi/models/unet_blocks.py
pointer only (licence: NONE) · get_code("b371191d08e93feb")

Repositories linked to this paper

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

Abstract

We introduce a new task -language-driven video inpainting, which uses natural language instructions to guide the inpainting process. This approach overcomes the limitations of traditional video inpainting methods that depend on manually labeled binary masks, a process often tedious and labor-intensive. We present the Remove Objects from Videos by Instructions (ROVI) dataset, containing 5,650 videos and 9,091 inpainting results, to support training and evaluation for this task. We also propose a novel diffusion-based language-driven video inpainting framework, the first end-to-end baseline for this task, integrating Multimodal Large Language Models to understand and execute complex language-based inpainting requests effectively. Our comprehensive results showcase the dataset's versatility and the model's effectiveness in various language-instructed inpainting scenarios. We have made datasets, code, and models publicly available at https://github.com/jianzongwu/Language-Driven-Video-Inpainting.

For agents

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

get_harvested_code_for_paper("2401.10226")
get_code_for_paper("2401.10226")
have("2401.10226")

Connect an agent — have() is free.