SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2409.04847 · NeurIPS · 2024

Rethinking The Training And Evaluation of Rich-Context Layout-to-Image Generation

Zheng Zhang, Tianjun Xiao, Tong He, Jiaxin Cheng, Zixu Zhao, Yicong Zhou

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 13 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
cplusx/rich_context_l2i canonical 10 of 13
FunctionStatusWhere it lives
clip_image Ran cplusx/rich_context_l2i/callbacks/common.py
pointer only (licence: NONE) · get_code("43749669b9801ca6")
exists Ran cplusx/rich_context_l2i/misc_utils/model_utils.py
pointer only (licence: NONE) · get_code("608e364a9d2376a3")
format_dtype_and_shape Ran cplusx/rich_context_l2i/callbacks/common.py
pointer only (licence: NONE) · get_code("43cc14973f357339")
get_coco_id_mapping Ran cplusx/rich_context_l2i/diffusion_dataset/COCOStuff.py
pointer only (licence: NONE) · get_code("862ce6b655ad9e1f")
get_obj_from_str Ran cplusx/rich_context_l2i/misc_utils/model_utils.py
pointer only (licence: NONE) · get_code("221b2d116fdf1032")
load_boxdiff_image Ran cplusx/rich_context_l2i/evaluation.py
pointer only (licence: NONE) · get_code("47ed05cc10ac81e2")
load_gligen_image Ran cplusx/rich_context_l2i/evaluation.py
pointer only (licence: NONE) · get_code("59fdc706e1142dc3")
load_instdiff_image Ran cplusx/rich_context_l2i/evaluation.py
pointer only (licence: NONE) · get_code("8e721c9b6023bbf1")
rename_duplicated_label_with_index Ran cplusx/rich_context_l2i/diffusion_dataset/LAION_synthetic.py
pointer only (licence: NONE) · get_code("41527b34aedeb1bb")
unnorm Ran cplusx/rich_context_l2i/callbacks/common.py
pointer only (licence: NONE) · get_code("5639833613f1f408")
get_cocostuff_caption_dataset Not yet run cplusx/rich_context_l2i/diffusion_dataset/COCOStuff.py
pointer only (licence: NONE) · get_code("cf2e00fe902bb93e")
get_cocostuff_dataset Not yet run cplusx/rich_context_l2i/diffusion_dataset/COCOStuff.py
pointer only (licence: NONE) · get_code("b8024f9ffe4b18f6")
instantiate_from_config Not yet run cplusx/rich_context_l2i/misc_utils/model_utils.py
pointer only (licence: NONE) · get_code("f0ac78b4e05be194")

Repositories linked to this paper

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

Abstract

Figure 1: The proposed method demonstrates the ability to accurately generate objects with complex descriptions in the correct locations while faithfully preserving the details specified in the text. In contrast, existing methods such as BoxDiff [57], R&B [56], GLIGEN [25], and InstDiff [54] struggle with the complex object descriptions, leading to errors in the generated objects.

For agents

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

get_harvested_code_for_paper("2409.04847")
get_code_for_paper("2409.04847")
have("2409.04847")

Connect an agent — have() is free.