20 results for “FastEEP”
CS papers onlyHybrid search: Keyword + semantic, ranked by combined score.ⓘ
Want pure semantic search? Try claim verification →
Yunfei Feng, Jingkai He, Shiqi Liu, Cheng Zhang +2 more
The paper presents FastEEP, a system for fast intra-instance scaling of Mixture-of-Experts (MoE) models in production.
The paper argues that the actor model is suitable for high-frequency trading and presents a C++20 framework, kaspar-hft, that adapts the model for HFT.
This paper implements and measures the cost of running a streaming speech enhancer on a CPU, achieving a 3.3x speedup compared to the fp32 ONNX Runtime graph.
This paper proposes a lightweight encoder-based MEL solution called FAST-MEL that meets three objectives: high linking accuracy, computational efficiency, and storage efficiency.
Bin Ma, Min Yu, Vijay Kodamalla, Jialin Liu +1 more
The authors present Difflet, a serving engine that accelerates diffusion model denoiser steps by fusing attention and reusing earlier outputs.
SOCI (Seekable OCI) is a lazy-loading architecture that reduces container image pulling time in Kubernetes environments by building an index over standard OCI images and serving file accesses via HTTP…
MimeLens is a novel, position-agnostic BERT-style encoder that accurately detects file types from arbitrary binary fragments, outperforming existing methods like Magika, especially on non-standard inp…
A benchmark called GenOlympiad is introduced to evaluate the performance of text-to-image models on challenging tasks.
Ray A. O. Sinurat, Yuyang Huang, Nanqinqin Li, Mark Powers +3 more
The authors introduce StorRep, a set of extensible and reproducible storage experiment artifacts.
Evaluates cost-aware abstention for IT support ticket routing using large language models and traditional classifiers.
Ashkan Pirmani, Ilse Vermeulen, Goran Vinterhalter, Lotte Geys +6 more
FLKit is an open, community-maintained onboarding toolkit for federated learning that provides a structured place for multidisciplinary teams to begin federated learning projects.
The study demonstrates that LLMs exhibit significant, language-driven disparities in medical triage recommendations, recommending emergency care more frequently for English and Arabic prompts, even wh…
This paper proposes TALLOP, a method that jointly optimizes local training workload and communication compression under per-round time constraints in federated learning.
Ke Zhang, Xiaoyang Yu, Haoyu Li, Shuai Wang +2 more
The paper introduces WeSep, a unified framework for Target Speaker Extraction that decouples cue modules and separator backbones, enabling flexible cue injection and integration of diverse modalities.
This paper investigates the instability of GEPA's system-aware merge operator across different models and datasets.
Wei Zheng, Yang Yan, Yiyang Shao, Jinyang Li +5 more
The paper proposes A2X, an LLM-native progressive-disclosure scheme that structures service taxonomies hierarchically and searches them layer-by-layer at query time, solving context overflow and impro…
Yihui Zhang, Tianyu Wo, Jinghao Wang, Xiaoyang Sun +6 more
This paper presents SpecBox, a runtime system for LLM agents that uses speculative sandbox preallocation to improve resource utilization and reduce interactive tail latency.