Institution profile

Hugging Face

Industry researcheurope · fr
Official website
Research library91linked papers
Opportunities3open roles
Selected work

Representative Papers

SmolVLM: Redefining small and efficient multimodal models

Apr 07, 2025

To address the high GPU memory consumption and low inference efficiency of large vision-language models (VLMs) on mobile and edge devices, this work proposes an architecture–tokenization–data co-optimization paradigm tailored for resource-constrained scenarios. Methodologically, we design a lightweight Transformer backbone, introduce sparse image/video tokenization strategies, and construct a high-quality, compact multimodal dataset trained via curriculum learning. Our key contributions are: (1) SmolVLM-256M achieves <1 GB GPU memory usage during inference while outperforming Idefics-80B in accuracy; (2) the 2.2B-parameter variant attains state-of-the-art performance on both image and video understanding tasks with significantly lower memory footprint; and (3) this is the first demonstration of a small-parameter VLM systematically surpassing ultra-large models across multimodal understanding benchmarks—establishing a new paradigm for efficient VLM deployment on edge devices.

2 citationsRead paper

DynEval: Holistic Evaluations of T2I Generative Models in the Wild

Jul 13, 2026

This work addresses the challenge that existing automatic evaluation methods for text-to-image (T2I) generation models struggle to accurately identify fine-grained failure modes—such as semantic mismatches and compositional errors—at scale. To this end, the authors propose DynEval, a dynamic evaluation framework that introduces a structured pipeline to decouple assessment of text-image alignment from image quality. They further develop two novel datasets, GenDB and DynEvalInstruct, via a hierarchical prompt-and-model generation strategy. Leveraging knowledge distillation and curriculum learning, DynEval enables a compact evaluator to efficiently inherit capabilities from large models through full-parameter fine-tuning. Experiments demonstrate that DynEval significantly outperforms current evaluators across 11 benchmarks, achieves higher correlation with human judgments, and provides the first fine-grained analysis of 36 T2I models across 42 subclasses and 9 semantic dimensions.

0 citationsRead paper
Recent publications

Latest Papers

DynEval: Holistic Evaluations of T2I Generative Models in the Wild

Jul 13, 2026

This work addresses the challenge that existing automatic evaluation methods for text-to-image (T2I) generation models struggle to accurately identify fine-grained failure modes—such as semantic mismatches and compositional errors—at scale. To this end, the authors propose DynEval, a dynamic evaluation framework that introduces a structured pipeline to decouple assessment of text-image alignment from image quality. They further develop two novel datasets, GenDB and DynEvalInstruct, via a hierarchical prompt-and-model generation strategy. Leveraging knowledge distillation and curriculum learning, DynEval enables a compact evaluator to efficiently inherit capabilities from large models through full-parameter fine-tuning. Experiments demonstrate that DynEval significantly outperforms current evaluators across 11 benchmarks, achieves higher correlation with human judgments, and provides the first fine-grained analysis of 36 T2I models across 42 subclasses and 9 semantic dimensions.

0 citationsRead paper

Flash-BoN: Instant Drafts for Inference-Time Scaling in Diffusion Models

Jul 05, 2026

Existing diffusion model inference acceleration methods overly rely on intermediate verification, neglecting the computational cost of generation itself and thereby distorting efficiency evaluations. This work proposes Flash-BoN, a framework that unifies for the first time three acceleration strategies—timestep truncation, network layer skipping, and activation surrogation—and introduces a low-cost draft candidate pool coupled with a multi-stage verification mechanism. Under a fixed wall-clock time budget, Flash-BoN prioritizes breadth-wise exploration over frequent validation. The method consistently outperforms existing approaches across three benchmarks and three model scales, achieving an 8% AUC improvement in large models. When combined with orthogonal techniques such as prompt optimization, it yields a 16% AUC gain, while also revealing the critical role of candidate diversity in generation quality and convergence of reinforcement learning-based post-training.

0 citationsRead paper