arXiv Machine Learning
HydraLLM: Adaptive Width Slicing for Elastic Text and Vision-Language Models
Read the original on arXiv Machine Learning →The Flow has not summarised this story yet — read it at arXiv Machine Learning.
Related stories
arXiv AI
Oct 7
Stepped MoE: Segment-Level Routing with Configurable Inference Complexity
arXiv:2610.07348v1 Announce Type: cross Abstract: Training large language models (LLMs) is resource-intensive, and adapting them for diverse deployment scenarios with varying computational constraint...
arXiv AI
Jul 28
cMoLLM at Scale: Horizontal Scaling Laws for Mixture-of-LLMs
arXiv:2607. 22577v1 Announce Type: new Abstract: Scaling large language models (LLMs) has driven their success, yet dense Transformers couple capacity and computation: every parameter is activated for every token, making training and inference costs grow linearly with model size-a critical bottleneck as models approach trillion-parameter regimes.
Hugging Face Trending Papers
Aug 4
ParVL: Parallel Scaling and Expandable Compute Allocation for Multimodal LLMs
Existing scaling strategies for Multimodal Large Language Models (MLLMs) typically expand either model parameters or sequential inference computation, incurring substantial memory or latency overhead. More importantly, most existing methods fail to alter the rigid, fixed computation allocation between the Vision Transformer and the Large Language Model components, limiting task-specific optimization.
arXiv Machine Learning
Sep 21
Elastic Threshold Attention: Learned Contextual Sparsity for Long-Context Decoding
Elastic Threshold Attention (ETA) is a trainable attention mechanism that dynamically predicts contextual thresholds from query representations, enabling selective pruning of KV cache tokens during long‑context decoding. By multiplicatively suppressing sub‑threshold logits during training, ETA avoids representation collapse and eliminates localized attention sinks, allowing a 1.45B model to match dense attention performance at roughly 85% training sparsity and 38% active decode density. At inference, a custom Triton kernel achieves up to 2.5× faster decoding on sequences up to 512K tokens, and an offline calibration step can further reduce compute by 27% by freezing per‑head thresholds.
arXiv AI
Aug 25
What Transfers from Text to Vision? Capability Scaling Laws and Transfer Dynamics for VLMs
The paper introduces the Capability-Driven Multimodal Scaling Law, a cross-family framework that predicts vision-language model (VLM) benchmark accuracy from a low-dimensional textual capability score extracted via PCA. By training over 150 VLMs on 34 large language models across seven families, the authors demonstrate that the law accurately extrapolates transfer rates from 8B to 72B‑parameter backbones, predicts full training trajectories, and generalizes to unseen model families. The study also reveals actionable insights, such as certain textual benchmarks negatively correlating with multimodal performance and base LLMs outperforming instruction-tuned counterparts as VLM backbones due to higher absorption rates.
arXiv AI
Oct 7
VisionWeave: Weaving Elastic Visual Representations as a Native Capability of MLLMs
VisionWeave introduces elastic visual representation weaving, a native capability for multimodal large language models that learns where and at what granularity to encode visual information. The method combines a gated spatial pooler for coarse representations with a granularity router that allocates content‑adaptive token usage, trained end‑to‑end on large‑scale data. Experiments on Qwen3.5‑4B and Qwen3.8‑27B show that VisionWeave can save 43.0% of tokens while preserving 98.9% of performance across eight benchmarks, and delivers significant throughput gains and latency reductions when deployed on the SGLang serving engine.