Kim Hojin

논문은 아니고 메모

Direct Preference Optimization: Your Language Model is Secretly a Reward Model을 읽고

SFT Memorizes, RL Generalizes: A Comparative Study of Foundation Model Post-training 논문 리뷰

HART : EFFICIENT VISUAL GENERATION WITH HYBRID AUTOREGRESSIVE TRANSFORMER 논문 리뷰

DIFFUSION MODELS ARE REAL-TIME GAME ENGINES

ONE STEP DIFFUSION VIA SHORTCUT MODELS 논문 리뷰

Common Diffusion Noise Schedules and Sample Steps are Flawed 논문 리뷰

SANA by NVIDIA paper review 논문 리뷰

DAC를 읽었다.

High-Resolution Image Synthesis with Latent Diffusion Models 논문 읽기

stable audio 논문 paper 리뷰

animateDiff 논문 paper

GAN 읽기

InstantDrag: Improving Interactivity in Drag-based Image Editing 논문 리뷰

SimCSE 논문 리뷰

Positive, Negative and Neutral: Modeling Implicit Feedback in Session-based News Recommendation paper review, 논문 리뷰

Autoregressive Image Generation without Vector Quantization 논문 리뷰

cycle3D 논문 리뷰

LRM 3d paper review

Efficient Geometry-aware 3D Generative Adversarial Networks 리뷰 for tri-plane understanding

3DGS 논문 리뷰

MAGVIT2 논문 리뷰

videopoet 논문 리뷰

Generative Modeling by Estimating Gradients of the Data Distribution 양송 블로그 리뷰

VASA-1 논문 리뷰

Rich Human Feedback for Text-to-Image Generation 논문 리뷰

The Platonic Representation Hypothesis 논문 리뷰

ViVid-1-to-3 논문 리뷰

VAR paper

Min-SNR-감마 : 논문 리뷰

논문은 아니고, RAG 관련 글에 대한 번역

[paper] LLaVA : Visual Instruction Tuning

[논문] Multi-persona LLM, UNLEASHING COGNITIVE SYNERGY IN LARGE LANGUAGE MODELS: A TASK SOLVING AGENT THROUGH MULTI-PERSONA SELF-COLLABORATION

사고방식

LLM을 Inference할 때 최적화하는 방법에 관하여

About suffering

Patterns for Building LLM-based Systems이라는 긴 아티클을 읽고 중요하다고 생각되는 부분만 뽑아서 정리한 내용

LoRA에 관해서. 읽고 간단하게 작성 논문 : https://arxiv.org/abs/2106.09685