126 citations · 376 across the 37 of their papers we have counts for
39 papers
Embarrassingly Simple Self-Distillation Improves Code Generation
Ruixiang Zhang, Richard He Bai, Huangjie Zheng +3
Can a large language model (LLM) improve at code generation using only its own raw outputs, without a verifier, a teacher model, or reinforcement learning? We answer in the affirma…
Rejection Mixing: Fast Semantic Propagation of Mask Tokens for Efficient DLLM Inference
Yushi Ye, Feng Hong, Huangjie Zheng +4
Diffusion Large Language Models (DLLMs) promise fast non-autoregressive inference but suffer a severe quality-speed trade-off in parallel decoding. This stems from the ''combinator…
Continuously Augmented Discrete Diffusion model for Categorical Generative Modeling
Huangjie Zheng, Shansan Gong, Ruixiang Zhang +5
Standard discrete diffusion models treat all unobserved states identically by mapping them to an absorbing [MASK] token. This creates an 'information void' where semantic informati…
Score Distillation of Flow Matching Models
Mingyuan Zhou, Yi Gu, Huangjie Zheng +5
Diffusion models achieve high-quality image generation but are limited by slow iterative sampling. Distillation methods alleviate this by enabling one- or few-step generation. Flow…
ChipChat: Low-Latency Cascaded Conversational Agent in MLX
Tatiana Likhomanenko, Richard He Bai, Zijin Gu +8
The emergence of large language models (LLMs) has transformed spoken dialog systems, yet the optimal architecture for real-time on-device voice agents remains an open question. Whi…
Wide-In, Narrow-Out: Revokable Decoding for Efficient and Effective DLLMs
Feng Hong, Geng Yu, Yushi Ye +5
Diffusion Large Language Models (DLLMs) have emerged as a compelling alternative to Autoregressive models, designed for fast parallel generation. However, existing DLLMs are plague…
Flexible Language Modeling in Continuous Space with Transformer-based Autoregressive Flows
Ruixiang Zhang, Shuangfei Zhai, Jiatao Gu +6
Autoregressive models have driven remarkable progress in language modeling. Their foundational reliance on discrete tokens, unidirectional context, and single-pass decoding, while…
DiffuCoder: Understanding and Improving Masked Diffusion Models for Code Generation
Shansan Gong, Ruixiang Zhang, Huangjie Zheng +4
Diffusion large language models (dLLMs) are compelling alternatives to autoregressive (AR) models because their denoising models operate over the entire sequence. The global planni…
STARFlow: Scaling Latent Normalizing Flows for High-resolution Image Synthesis
Jiatao Gu, Tianrong Chen, David Berthelot +7
We present STARFlow, a scalable generative model based on normalizing flows that achieves strong performance in high-resolution image synthesis. The core of STARFlow is Transformer…
TADA: Improved Diffusion Sampling with Training-free Augmented Dynamics
Tianrong Chen, Huangjie Zheng, David Berthelot +3
Diffusion models have demonstrated exceptional capabilities in generating high-fidelity images but typically suffer from inefficient sampling. Many solver designs and noise schedul…
Normalizing Flows are Capable Generative Models
Shuangfei Zhai, Ruixiang Zhang, Preetum Nakkiran +7
Normalizing Flows (NFs) are likelihood-based models for continuous inputs. They have demonstrated promising results on both density estimation and generative modeling tasks, but ha…
What If We Recaption Billions of Web Images with LLaMA-3?
Xianhang Li, Haoqin Tu, Mude Hui +9
Web-crawled image-text pairs are inherently noisy. Prior studies demonstrate that semantically aligning and enriching textual descriptions of these pairs can significantly enhance…
Guided Score identity Distillation for Data-Free One-Step Text-to-Image Generation
Mingyuan Zhou, Zhendong Wang, Huangjie Zheng +1
Diffusion-based text-to-image generation models trained on extensive text-image pairs have demonstrated the ability to produce photorealistic images aligned with textual descriptio…
Score identity Distillation: Exponentially Fast Distillation of Pretrained Diffusion Models for One-Step Generation
Mingyuan Zhou, Huangjie Zheng, Zhendong Wang +2
We introduce Score identity Distillation (SiD), an innovative data-free method that distills the generative capabilities of pretrained diffusion models into a single-step generator…
Learning Stackable and Skippable LEGO Bricks for Efficient, Reconfigurable, and Variable-Resolution Diffusion Modeling
Huangjie Zheng, Zhendong Wang, Jianbo Yuan +5
Diffusion models excel at generating photo-realistic images but come with significant computational costs in both training and sampling. While various techniques address these comp…
Beta Diffusion
Mingyuan Zhou, Tianqi Chen, Zhendong Wang +1
We introduce beta diffusion, a novel generative modeling method that integrates demasking and denoising to generate data within bounded ranges. Using scaled and shifted beta distri…
Class-Balancing Diffusion Models
Yiming Qin, Huangjie Zheng, Jiangchao Yao +2
Diffusion-based models have shown the merits of generating high-quality visual data while preserving better diversity in recent studies. However, such observation is only justified…
POUF: Prompt-oriented unsupervised fine-tuning for large pre-trained models
Korawat Tanwisuth, Shujian Zhang, Huangjie Zheng +2
Through prompting, large-scale pre-trained models have become more expressive and powerful, gaining significant attention in recent years. Though these big models have zero-shot ca…
Re-imagine the Negative Prompt Algorithm: Transform 2D Diffusion into 3D, alleviate Janus problem and Beyond
Mohammadreza Armandpour, Ali Sadeghian, Huangjie Zheng +2
Although text-to-image diffusion models have made significant strides in generating images from text, they are sometimes more inclined to generate images like the data on which the…
Patch Diffusion: Faster and More Data-Efficient Training of Diffusion Models
Zhendong Wang, Yifan Jiang, Huangjie Zheng +5
Diffusion models are powerful, but they require a lot of time and data to train. We propose Patch Diffusion, a generic patch-wise training framework, to significantly reduce the tr…
DR2: Diffusion-based Robust Degradation Remover for Blind Face Restoration
Zhixin Wang, Xiaoyun Zhang, Ziying Zhang +4
Blind face restoration usually synthesizes degraded low-quality data with a pre-defined degradation model for training, while more complex cases could happen in the real world. Thi…
CARD: Classification and Regression Diffusion Models
Xizewen Han, Huangjie Zheng, Mingyuan Zhou
Learning the distribution of a continuous or categorical response variable given its covariates is a fundamental problem in statistics and machine l…
Diffusion-GAN: Training GANs with Diffusion
Zhendong Wang, Huangjie Zheng, Pengcheng He +2
Generative adversarial networks (GANs) are challenging to train stably, and a promising remedy of injecting instance noise into the discriminator input has not been very effective…
Representing Mixtures of Word Embeddings with Mixtures of Topic Embeddings
Dongsheng Wang, Dandan Guo, He Zhao +4
A topic model is often formulated as a generative model that explains how each word of a document is generated given a set of topics and document-specific topic proportions. It is…