V
主页
ZipLoRA: Any Subject in Any Style by Effectively Merging LoRAs
发布人
【加群】 一起来刷arxiv,请加vx: pwbot02(请备注:b站arxiv) 【彩蛋】 可以试试/ask + 你的提问和本篇论文进行交流 【论文标题】 ZipLoRA: Any Subject in Any Style by Effectively Merging LoRAs 【论文简述】 该论文介绍了一种名为ZipLoRA的方法,可以廉价且有效地合并独立训练的风格和主题适应模型,实现用户提供的主题和风格的生成。通过对多种主题和风格的组合进行实验,结果表明ZipLoRA能够产生令人信服的结果,并在主题和风格保真度方面相比基线方法有显著改进,同时保持了重新语境化能力。 【引导阅读的问题】 如何在保证主题和风格保真度的同时,合并独立训练的风格和主题适应模型? 【论文链接】 https://arxiv.org/pdf/2311.13600
打开封面
下载高清视频
观看高清视频
视频下载器
MDPI惨遭“降级”,拒稿率高达88%!没实力如何3个月逆袭Q2?
FluxMusic: 基于FLUX的高效文本到音乐生成系统
Distil-Whisper: Robust Knowledge Distillation via Large-Scale Pseudo Labelling
VideoCon: Robust Video-Language Alignment via Contrast Captions
TXYZ怎么用?
In-Context Learning Creates Task Vectors
GAN生成高保真3D几何:渲染每个像素以提升3D GAN质量
港科大Sherry教授课题组组会(全英有字幕):如何高效用LaTex写学术论文
PhysGaussian: Physics-Integrated 3D Gaussians for Generative Dynamics
Text-to-Sticker: Style Tailoring Latent Diffusion Models for Human Expression
SOTOPIA-$π$: Interactive Learning of Socially Intelligent Language Agents
ReFT: Reasoning with Reinforced Fine-Tuning
VeRA: Vector-based Random Matrix Adaptation
Efficient Tool Use with Chain-of-Abstraction Reasoning
Tied-Lora: Enhacing parameter efficiency of LoRA with weight tying
中科院开源AI学术论文神器爆火!GPT学术版
Farzi Data: Autoregressive Data Distillation
Moral Foundations of Large Language Models
2024年学生党必备的网站,中科院开源的AI神器,GPT学术优化工具#学生党#AI神器#学术论文#中科院
单张图像到3D的高效生成:基于分摊生成的3D高斯模型
Points2Plans:从点云到具有可组合关系动力学的长期规划
【AI做PPT】通义千问免费AI做PPT神器火爆上线!效率暴增百倍!
地表最强AI写代码工具——Cursor:人人都是P8程序员
Understanding prompt engineering may not require rethinking generalization
【你绝对不知道的】这个网站藏着科研圈的宝藏,错过就太可惜了!
教女儿Python,骂哭了,狗狗保护她。
PALP: Prompt Aligned Personalization of Text-to-Image Models
Language Models can be Logical Solvers
Can Large Language Models be Good Path Planners? A Benchmark and Investigation o
TinyLlama:开源小型语言模型的新突破
基于3D高斯展开的实时少样本视角合成
Let's Synthesize Step by Step: Iterative Dataset Synthesis with Large Language M
RLVF: Learning from Verbal Feedback without Overgeneralization
An Image is Worth Multiple Words: Learning Object Level Concepts using Multi-Con
SANeRF-HQ:基于提示的高质量NeRF三维物体分割
Approximating Two-Layer Feedforward Networks for Efficient Transformers
Domain Generalization Guided by Gradient Signal to Noise Ratio of Parameters
Two-Stage Statistics-Aware Transformation for Style Transfer
SHINOBI: Shape and Illumination using Neural Object Decomposition via BRDF Optim
Label Supervised LLaMA Finetuning