SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2503.18446 · CVPR · 2025

Latent Space Super-Resolution for Higher-Resolution Image Generation with Diffusion Models

Seon Kim, Jinwoo Kim, Jinho Jeong, Sangmin Han

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 8 functions out of this paper's own repositories and ran 5 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
3587jjh/lsrna — 5 of 8
FunctionStatusWhere it lives
PatchEmbed Ran 3587jjh/lsrna/lsr/swinir.py
pointer only (licence: NONE) · get_code("ff70504409e0ef02")
PatchUnEmbed Ran 3587jjh/lsrna/lsr/swinir.py
pointer only (licence: NONE) · get_code("223f9228b44f77f4")
RSTB Ran 3587jjh/lsrna/lsr/swinir.py
pointer only (licence: NONE) · get_code("64564d9d8d7ee992")
UpsampleOneStep Ran 3587jjh/lsrna/lsr/swinir.py
pointer only (licence: NONE) · get_code("2954dcfb57741b89")
WindowAttention Ran 3587jjh/lsrna/lsr/swinir.py
pointer only (licence: NONE) · get_code("c03392ec1f6065d1")
BasicLayer Not yet run 3587jjh/lsrna/lsr/swinir.py
pointer only (licence: NONE) · get_code("bf46e677688a82bb")
SwinIR Not yet run 3587jjh/lsrna/lsr/swinir.py
pointer only (licence: NONE) · get_code("0dd67a76da7dee76")
SwinTransformerBlock Not yet run 3587jjh/lsrna/lsr/swinir.py
pointer only (licence: NONE) · get_code("7d675ffafa9e4051")

Repositories linked to this paper

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

Abstract

Pixelsmith LSRNA-Pixelsmith (Ours) 505 sec (100%) 314 sec (62%) DemoFusion LSRNA-DemoFusion (Ours) 1507 sec (100%) 506 sec (34%) Zoomed-in view Zoomed-in view Zoomed-in view Zoomed-in view 16× Resolution [4 × 4096] 16× Resolution [4 × 4096] 1× 16× Resolution [4 × 4096] 16× Resolution [4 × 4096] 1× Figure 1. Comparisons of 16× image generation with and without LSRNA framework. Our proposed LSRNA framework improves reference-based higher-resolution image generation, enhancing detail and sharpness beyond the native resolution of SDXL [36] (1024 2 ) while achieving faster generation speeds.

For agents

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

get_harvested_code_for_paper("2503.18446")
get_code_for_paper("2503.18446")
have("2503.18446")

Connect an agent — have() is free.