
微调 8
-
论文笔记《Real-Time Video Inference on Edge Devices via Adaptive Model Streaming》
论文 - 《Real-Time Video Inference on Edge Devices via Adaptive Model Streaming》 代码 - Github 关键词 - 实时视频推理、边缘智能、蒸馏、端云协作、适应、ICCV2021 摘要 研究问题 在移动电话和无人机等边缘设备
-
论文笔记《Unsupervised Domain Adaptive Visual Question Answering in the era of MLLMs》
论文笔记 - 《Unsupervised Domain Adaptive Visual Question Answering in the era of Multi-modal Large Language Models》 关键词 - 问答、特征对齐、多模态、域适应、WACV2025 1 介绍 研究
-
论文笔记《Cloud-Device Collaborative Learning for Multimodal Large Language Models》
论文-《Cloud-Device Collaborative Learning for Multimodal Large Language Models》 关键词:云端-设备协作、多模态、大模型、CVPR2024 摘要 问题背景:多模态大语言模型(MLLMs)在图像描述生成、常识推理和视觉场景理解等
-
论文笔记《Task-Oriented Feature Compression for Multimodal Understanding via Device-Edge Co-Inference》
论文:《Task-Oriented Feature Compression for Multimodal Understanding via Device-Edge Co-Inference》 摘要 研究背景 大多数LMM (large multimodal models) 的推理请求来自边缘设备,
-
论文笔记《Self-Adapting Large Visual-Language Models to Edge Devices...》
论文-《Self-Adapting Large Visual-Language Models to Edge Devices across Visual Modalities》 代码-Github 摘要 研究问题 视觉-语言(VL)模型的进展引发了对其在边缘设备上部署的兴趣,但在处理多样化视觉模态、
-
论文笔记《Towards Efficient Asynchronous Federated Learning in Heterogeneous Edge Environments》
论文地址-《Towards Efficient Asynchronous Federated Learning in Heterogeneous Edge Environments》 摘要 研究解决的问题 边缘设备通常具有异构的计算能力和数据分布,阻碍了协同训练的效率。现有的工作开发了陈旧感知的半异
-
论文笔记《Agglomerative Federated Learning: Empowering Larger...》
论文地址-《Agglomerative Federated Learning: Empowering Larger Model Training via End-Edge-Cloud Collaboration》 摘要 尽管分层联邦学习HFL支持适合EECC的多层模型聚合,先前的工作在所有计算节点上
-
论文笔记《ZeRO++: Extremely Efficient Collective Communication for Giant Model Training》
论文地址-《ZeRO++: Extremely Efficient Collective Communication for Giant Model Training》 摘要 ZeRO的缺点: 当在低带宽集群上进行训练时,或者在规模上迫使每个 GPU 的批量大小变小时,ZeRO 的有效吞吐量受到限制