SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2403.05049 · ECCV · 2024

XPSR: Cross-modal Priors for Diffusion-based Image Super-Resolution

Ming Sun, Kun Yuan, Kai Zhao, Chao Zhou, Jinhua Hao, Yunpeng Qu, Qizhi Xie

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 20 functions out of this paper's own repositories and ran 9 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
qyp2000/XPSR canonical 9 of 20
FunctionStatusWhere it lives
attn_forward_new_pt2_0 Ran qyp2000/XPSR/myutils/vaehook.py
pointer only (licence: NONE) · get_code("177c1cf8cd11be39")
convert_image_to_fn Ran qyp2000/XPSR/myutils/img_util.py
pointer only (licence: NONE) · get_code("1b81186cfb3f8848")
exists Ran qyp2000/XPSR/dataloader/localdataset.py
pointer only (licence: NONE) · get_code("608e364a9d2376a3")
identity Ran qyp2000/XPSR/myutils/misc.py
pointer only (licence: NONE) · get_code("9910e2fc297f8665")
inplace_nonlinearity Ran qyp2000/XPSR/myutils/vaehook.py
pointer only (licence: NONE) · get_code("76a017e88938061d")
rand_name Ran qyp2000/XPSR/myutils/misc.py
pointer only (licence: NONE) · get_code("bdd74d6bfd709ddf")
save_videos_grid Ran qyp2000/XPSR/myutils/img_util.py
pointer only (licence: NONE) · get_code("73408b539d7e8eca")
window_partition Ran qyp2000/XPSR/models/xpsr/unet_2d_blocks.py
pointer only (licence: NONE) · get_code("bf260685de569a29")
window_reverse Ran qyp2000/XPSR/models/xpsr/unet_2d_blocks.py
pointer only (licence: NONE) · get_code("184db8216d8a8a05")
attn_forward_new Not yet run qyp2000/XPSR/myutils/vaehook.py
pointer only (licence: NONE) · get_code("08822a10826682a0")
cond_cast_float Not yet run qyp2000/XPSR/myutils/devices.py
pointer only (licence: NONE) · get_code("a1316b96fc0dc988")
cond_cast_unet Not yet run qyp2000/XPSR/myutils/devices.py
pointer only (licence: NONE) · get_code("6e1abd17fa1ce3b2")
convert_lora Not yet run qyp2000/XPSR/myutils/convert_lora_safetensor_to_diffusers.py
pointer only (licence: NONE) · get_code("35fa1e7fbadd8f3d")
get_device_for Not yet run qyp2000/XPSR/myutils/devices.py
pointer only (licence: NONE) · get_code("a861df384ac42eeb")
mesh_grid Not yet run qyp2000/XPSR/dataloader/degradations.py
pointer only (licence: NONE) · get_code("794d4f7fc184e7e5")
opt_parse Not yet run qyp2000/XPSR/dataloader/realesrgan.py
pointer only (licence: NONE) · get_code("fce64ea4638101fc")
pdf2 Not yet run qyp2000/XPSR/dataloader/degradations.py
pointer only (licence: NONE) · get_code("321b9c73d589eb9e")
sigma_matrix2 Not yet run qyp2000/XPSR/dataloader/degradations.py
pointer only (licence: NONE) · get_code("e3444ff2b9d24600")
window_process Not yet run qyp2000/XPSR/models/xpsr/unet_2d_blocks.py
pointer only (licence: NONE) · get_code("d40a55c91fa32c4f")
zero_module Not yet run qyp2000/XPSR/models/xpsr/controlnet.py
pointer only (licence: NONE) · get_code("da94debb8019ad46")

Repositories linked to this paper

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

Abstract

Diffusion-based methods, endowed with a formidable generative prior, have received increasing attention in Image Super-Resolution (ISR) recently. However, as low-resolution (LR) images often undergo severe degradation, it is challenging for ISR models to perceive the semantic and degradation information, resulting in restoration images with incorrect content or unrealistic artifacts. To address these issues, we propose a Cross-modal Priors for Super-Resolution (XPSR) framework. Within XPSR, to acquire precise and comprehensive semantic conditions for the diffusion model, cutting-edge Multimodal Large Language Models (MLLMs) are utilized. To facilitate better fusion of cross-modal priors, a Semantic-Fusion Attention is raised. To distill semantic-preserved information instead of undesired degradations, a Degradation-Free Constraint is attached between LR and its high-resolution (HR) counterpart. Quantitative and qualitative results show that XPSR is capable of generating high-fidelity and high-realism images across synthetic and real-world datasets. Codes are released at https://github.com/qyp2000/XPSR.

For agents

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

get_harvested_code_for_paper("2403.05049")
get_code_for_paper("2403.05049")
have("2403.05049")

Connect an agent — have() is free.