JyChen9811/FaithDiff
[CVPR 2025] FaithDiff for Classic Film Rejuvenation, Old Photo Revival, Social Media Restoration, Image Enhancement and AIGC Enhancement.
Leverages pre-trained diffusion models (SDXL) with semantic guidance from vision-language models (LLaVA) to preserve content fidelity during upscaling, using a two-stage training pipeline that conditions the diffusion process on image degradation and semantic understanding. Supports memory-efficient inference through FP8 quantization and CPU offloading, enabling 8K+ restoration on consumer GPUs, and integrates directly with Hugging Face Diffusers for seamless adoption in existing workflows.
240 stars.
Stars
240
Forks
15
Language
Python
License
MIT
Category
Last pushed
Feb 22, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/JyChen9811/FaithDiff"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
UCSC-VLAA/story-iter
[ICLR 2026] A Training-free Iterative Framework for Long Story Visualization
PaddlePaddle/PaddleMIX
Paddle Multimodal Integration and eXploration, supporting mainstream multi-modal tasks,...
keivalya/mini-vla
a minimal, beginner-friendly VLA to show how robot policies can fuse images, text, and states to...
adobe-research/custom-diffusion
Custom Diffusion: Multi-Concept Customization of Text-to-Image Diffusion (CVPR 2023)
byliutao/1Prompt1Story
🔥ICLR 2025 (Spotlight) One-Prompt-One-Story: Free-Lunch Consistent Text-to-Image Generation...