https://github.com/gfnorg/diffusion-finetuning
공식 코드Amortizing intractable inference in diffusion models for vision, language, and control
태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.
The development of the brain's blood supply in an embryo involves a complex process with several stages. Initially, there is a network of connections between the carotid and basilar artery systems that usually disappear during development. However, in some cases, these connections remain, leading to variations in the adult cerebral circulation. Around the 24th day of embryonic life (at the 3 mm embryonic stage), the internal carotid arteries (ICA) emerge. They are formed from a combination of the 3rd branchial arch arteries and the distal segments of the paired dorsal aortae. The ICA initially provides all the blood needed by the developing brain. As the brain grows, particularly the occipital region, brainstem, and cerebellum, the blood supply from the ICA becomes insufficient, prompting the development of the posterior circulation. At this stage (4-5 mm embryonic stage), the posterior circulation, specifically the hindbrain, is supplied by two longitudinal neural arteries. These arteries receive blood from several temporary connections (anastomoses) with the ICA, including: Trigeminal artery (TA), Otic artery (OA), Hypoglossal artery (HA), Proatlantal artery (ProA). These carotid-vertebrobasilar anastomoses act as temporary bridges, providing blood flow to the developing posterior circulation until the vertebral arteries (VA) and the basilar artery (BA) are fully formed.6
The BA forms between the 5-8 mm stage by the fusion of the longitudinal neural arteries. As the posterior communicating artery develops and connects with the distal BA, the TA, OA, and HA typically regress. Unlike the TA, OA, and HA, the ProA persists until the VA are fully developed. A segment of the ProA becomes incorporated into the V3 segment of the VA and the distal portions of the occipital artery. The VA develop between the 7-12 mm stage from transverse connections between cervical intersegmental arteries, starting with the ProA and progressing downwards to the 6th intersegmental artery. This 6th intersegmental artery eventually forms the origin of the adult VA from the subclavian artery. 7-8
Persistence of the HA into adulthood, the focus of this discussion, is a rare occurrence. It is the second most common persistent carotid-vertebrobasilar anastomosis, following the persistent trigeminal artery. The failure of the HA to regress during embryological development leads to PPHA. This condition can impact the normal blood flow dynamics in the brain and, in certain cases, can be associated with cerebrovascular issues like aneurysms or ischemic events.
관련 논문과 평가 결과에 연결된 구현입니다. 출처에서 확인된 저자 구현에는 공식 코드 표시가 있습니다.
Amortizing intractable inference in diffusion models for vision, language, and control
GreenStableYolo: Optimizing Inference Time and Image Quality of Text-to-Image Generation
FuseDream: Training-Free Text-to-Image Generation with Improved CLIP+GAN Space Optimization
InstaFlow: One Step is Enough for High-Quality Diffusion-Based Text-to-Image Generation
Proactive Agents for Multi-Turn Text-to-Image Generation Under Uncertainty
ViSAGe: A Global-Scale Analysis of Visual Stereotypes in Text-to-Image Generation
JetFormer: An Autoregressive Generative Model of Raw Images and Text
Cross-Modal Contrastive Learning for Text-to-Image Generation
TIAM -- A Metric for Evaluating Alignment in Text-to-Image Generation
Exact Diffusion Inversion via Bi-directional Integration Approximation
AudioToken: Adaptation of Text-Conditioned Diffusion Models for Audio-to-Image Generation
CreativeSynth: Cross-Art-Attention for Artistic Image Synthesis with Multimodal Diffusion
LLM Blueprint: Enabling Text-to-Image Generation with Complex and Detailed Prompts
Accelerating Text-to-Image Editing via Cache-Enabled Sparse Diffusion Inference
StackGAN: Text to Photo-realistic Image Synthesis with Stacked Generative Adversarial Networks
StackGAN++: Realistic Image Synthesis with Stacked Generative Adversarial Networks
Intelligent Grimm -- Open-ended Visual Storytelling via Latent Diffusion Models
ConceptExpress: Harnessing Diffusion Models for Single-image Unsupervised Concept Extraction
ViCo: Plug-and-play Visual Condition for Personalized Text-to-image Generation
Evaluating Image Hallucination in Text-to-Image Generation with Question-Answering
Multimodal LLM-Guided Semantic Correction in Text-to-Image Diffusion
HiDream-I1: A High-Efficient Image Generative Foundation Model with Sparse Diffusion Transformer
HiDream-I1: A High-Efficient Image Generative Foundation Model with Sparse Diffusion Transformer
Hierarchical Masked Autoregressive Models with Low-Resolution Token Pivots
FireFlow: Fast Inversion of Rectified Flow for Image Semantic Editing
출처: Papers with Code · CC BY-SA 4.0. 아카이브 자료와 새로 수집한 논문은 별도로 관리합니다.