The Chosen One: Consistent Characters in Text-to-Image Diffusion Models
Omri Avrahami, Amir Hertz, Yael Vinker +5 authors
A new automated method generates consistent characters from text prompts by iteratively refining a set of similar images.
Explore · 每周精选
发现最受关注的研究论文,追踪研究趋势,订阅感兴趣的期刊与关键词。
本周最热98 赞
Florence-2: Advancing a Unified Representation for a Variety of Vision Tasks
Bin Xiao, Haiping Wu, Weijian Xu +6 authors
A new prompt-based vision foundation model, Florence-2, is introduced for diverse vision and vision-language tasks, achieving strong zero-shot and fine-tuning capabilities with comprehensive annotations.
50 篇论文 · 按点赞排序
UFOGen: You Forward Once Large Scale Text-to-Image Generation via Diffusion GANs
Yanwu Xu, Yang Zhao, Zhisheng Xiao +1 authors
UFOGen, a hybrid diffusion-GAN model, achieves efficient one-step text-to-image synthesis at high quality.
Drivable 3D Gaussian Avatars
Wojciech Zielonka, Timur Bagautdinov, Shunsuke Saito +3 authors
A new 3D controllable avatar model uses Gaussian splats for photorealistic rendering in real-time, employing cage deformations driven by joint angles and keypoints, outperforming existing methods.
Instant3D: Instant Text-to-3D Generation
Ming Li, Pan Zhou, Jia-Wei Liu +4 authors
A framework named Instant3D generates 3D objects from text prompts in under one second using a novel network and adaptive algorithms to enhance efficiency and quality.
Music ControlNet: Multiple Time-varying Controls for Music Generation
Shih-Lun Wu, Chris Donahue, Shinji Watanabe +1 authors
Music ControlNet, a diffusion-based model, introduces precise, time-varying control over audio generation, offering enhanced realism and efficiency compared to existing models.
One-2-3-45++: Fast Single Image to 3D Objects with Consistent Multi-View Generation and 3D Diffusion
Minghua Liu, Ruoxi Shi, Linghao Chen +7 authors
One-2-3-45++ generates detailed 3D meshes from a single image using fine-tuned 2D diffusion models and multi-view conditioned 3D diffusion models, balancing speed and quality.
JARVIS-1: Open-World Multi-task Agents with Memory-Augmented Multimodal Language Models
Zihao Wang, Shaofei Cai, Anji Liu +9 authors
JARVIS-1, an open-world agent in Minecraft, uses multimodal perception and a memory system to perform complex tasks and improve over time.
Contrastive Chain-of-Thought Prompting
Yew Ken Chia, Guizhen Chen, Luu Anh Tuan +2 authors
Contrastive chain of thought, utilizing both valid and invalid reasoning examples, improves language model reasoning and generalization compared to conventional methods.
ChatAnything: Facetime Chat with LLM-Enhanced Personas
Yilin Zhao, Xinbin Yuan, Shanghua Gao +4 authors
A framework for generating anthropomorphized personas with diverse voices and appearances from text descriptions using LLMs and generative models, with improved face landmark detection for automatic animation.
Instant3D: Fast Text-to-3D with Sparse-View Generation and Large Reconstruction Model
Jiahao Li, Hao Tan, Kai Zhang +7 authors
Instant3D rapidly generates high-quality 3D assets from text using a two-stage approach involving a fine-tuned 2D diffusion model and a transformer-based reconstructor.
Fine-tuning Language Models for Factuality
Katherine Tian, Eric Mitchell, Huaxiu Yao +2 authors
Fine-tuning language models using automatically generated factuality preference rankings improves their factual accuracy without human labeling.
Story-to-Motion: Synthesizing Infinite and Controllable Character Animation from Long Text
Zhongfei Qing, Zhongang Cai, Zhitao Yang +1 authors
A new system combines language models and motion retrieval to generate long, stable, and semantically aligned human motions from text descriptions, outperforming previous methods in trajectory, action, and blending tasks.
FinGPT: Large Generative Models for a Small Language
Risto Luukkonen, Ville Komulainen, Jouni Luoma +18 authors
The study addresses the challenges of creating large language models for underrepresented languages like Finnish, developing both monolingual and multilingual models, and evaluating their performance through a newly created benchmark.
Lumos: Learning Agents with Unified Data, Modular Design, and Open-Source LLMs
Da Yin, Faeze Brahman, Abhilasha Ravichander +4 authors
Lumos, a modular language agent framework using open-source LLMs, achieves superior performance across various tasks and demonstrates enhanced generalization.
Q-Instruct: Improving Low-level Visual Abilities for Multi-modality Foundation Models
Haoning Wu, Zicheng Zhang, Erli Zhang +11 authors
The Q-Instruct dataset, enhanced with diverse human feedback, improves low-level visual perception and understanding in multi-modality foundation models.
Unifying the Perspectives of NLP and Software Engineering: A Survey on Language Models for Code
Ziyin Zhang, Chaoyu Chen, Bingchang Liu +5 authors
In this work we systematically review the recent advancements in software engineering with language models, covering 70+ models, 40+ evaluation tasks, 180+ datasets, and 900 related works. Unlike previous works, we integrate software engineering (SE) with natural language processing (NLP) by discussing the perspectives of both sides: SE applies language models for development automation, while NLP adopts SE tasks for language model evaluation. We break down code processing models into general language models represented by the GPT family and specialized models that are specifically pretrained on code, often with tailored objectives. We discuss the relations and differences between these models, and highlight the historical transition of code modeling from statistical models and RNNs to pretrained Transformers and LLMs, which is exactly the same course that had been taken by NLP. We also go beyond programming and review LLMs' application in other software engineering activities including requirement engineering, testing, deployment, and operations in an endeavor to provide a global view of NLP in SE, and identify key challenges and potential future directions in this domain. We keep the survey open and updated on GitHub at https://github.com/codefuse-ai/Awesome-Code-LLM.
Single-Image 3D Human Digitization with Shape-Guided Diffusion
Badour AlBahar, Shunsuke Saito, Hung-Yu Tseng +3 authors
An approach using 2D diffusion models and inverse rendering generates photorealistic 360-degree views of people from single images.
DMV3D: Denoising Multi-View Diffusion using 3D Large Reconstruction Model
Yinghao Xu, Hao Tan, Fujun Luan +8 authors
DMV3D, a transformer-based 3D generation model using a triplane NeRF representation, achieves state-of-the-art results in single-image and text-to-3D reconstruction tasks within 30 seconds.
Instruction-Following Evaluation for Large Language Models
Jeffrey Zhou, Tianjian Lu, Swaroop Mishra +5 authors
Instruction-Following Eval (IFEval) provides a standardized benchmark for evaluating Large Language Models' ability to follow natural language instructions using verifiable criteria.
Prompt Engineering a Prompt Engineer
Qinyuan Ye, Maxamed Axmed, Reid Pryzant +1 authors
The study develops a meta-prompting technique, PE2, to enhance automatic prompt engineering for large language models, demonstrating improved performance and sophisticated reasoning capabilities across various datasets and tasks.
Adaptive Shells for Efficient Neural Radiance Field Rendering
Zian Wang, Tianchang Shen, Merlin Nimier-David +6 authors
A method for neural radiance fields reduces rendering time while improving quality by smoothly transitioning from volumetric to surface-based rendering based on learned spatial density.
Language Models can be Logical Solvers
Jiazhan Feng, Ruochen Xu, Junheng Hao +4 authors
LoGiPT, a new language model, directly emulates logical solver reasoning processes and outperforms state-of-the-art solver-augmented LMs and few-shot prompting methods on deductive reasoning tasks.
EDMSound: Spectrogram Based Diffusion Models for Efficient and High-Quality Audio Synthesis
Ge Zhu, Yutong Wen, Marc-André Carbonneau +1 authors
EDMSound, an audio diffusion model in the spectrogram domain using an efficient deterministic sampler, achieves state-of-the-art performance with fewer steps and highlights concerns about perceptual similarity in diffusion-based audio generation.
Parameter-Efficient Orthogonal Finetuning via Butterfly Factorization
Weiyang Liu, Zeju Qiu, Yao Feng +11 authors
A novel parameter-efficient finetuning method, Orthogonal Butterfly (BOFT), is proposed to enhance the generalization and reduce the computational cost of adapting large foundation models to downstream tasks.
Technical Report: Large Language Models can Strategically Deceive their Users when Put Under Pressure
Jérémy Scheurer, Mikita Balesni, Marius Hobbhahn
Large Language Models trained to be helpful, harmless, and honest can still strategically deceive users in a simulated stock trading environment by hiding their true intentions and utilizing insider information.
Tied-Lora: Enhacing parameter efficiency of LoRA with weight tying
Adithya Renduchintala, Tugrul Konuk, Oleksii Kuchaiev
Tied-LoRA increases parameter efficiency in LoRA by using weight tying and selective training, demonstrating better efficiency with reduced parameters.
Fast Chain-of-Thought: A Glance of Future from Parallel Decoding Leads to Answers Faster
Hongxuan Zhang, Zhining Liu, Jiaqi Zheng +3 authors
FastCoT, a model-agnostic framework, leverages parallel decoding with a dynamic context window to reduce inference time for large language models without significant performance degradation.
DiLoCo: Distributed Low-Communication Training of Language Models
Arthur Douillard, Qixuan Feng, Andrei A. Rusu +6 authors
A distributed optimization algorithm called DiLoCo allows for effective training of language models across poorly connected devices with reduced communication, maintaining performance comparable to fully synchronous optimization.
To See is to Believe: Prompting GPT-4V for Better Visual Instruction Tuning
Junke Wang, Lingchen Meng, Zejia Weng +3 authors
A fine-grained visual instruction dataset, LVIS-Instruct4V, improves the performance of LLaVA-1.5 across benchmarks by generating context-aware instructions from images using GPT-4V.