推荐开源项目：深入探索VQA v2的若干最新方法（2018）

推荐开源项目：深入探索VQA v2的若干最新方法（2018）

VQA2.0-Recent-Approachs-2018.pytorchA pytroch reimplementation of "Bilinear Attention Network", "Intra- and Inter-modality Attention", "Learning Conditioned Graph Structures", "Learning to count object", "Bottom-up top-down" for Visual Question Answering 2.0项目地址:https://gitcode.com/gh_mirrors/vq/VQA2.0-Recent-Approachs-2018.pytorch

在视觉问答（Visual Question Answering，简称VQA）领域，不断有新的研究推动着技术的边界。今天，我们聚焦于一个特别的开源项目——基于[Cyanogenoid/vqa-counting][0]的“Several Recent Approaches (2018) on VQA v2”，这是一个致力于实现和探索2018年多个视觉问答先进模型的宝藏库。

项目介绍

此项目重新实现了包括“Bottom-Up and Top-Down”注意力机制、双线性注意力网络、模态间与内在注意力、学习计数以及条件图结构学习

VQA2.0-Recent-Approachs-2018.pytorchA pytroch reimplementation of "Bilinear Attention Network", "Intra- and Inter-modality Attention", "Learning Conditioned Graph Structures", "Learning to count object", "Bottom-up top-down" for Visual Question Answering 2.0项目地址:https://gitcode.com/gh_mirrors/vq/VQA2.0-Recent-Approachs-2018.pytorch