<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[blueag1e的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/blueag1e</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; blueag1e]]></copyright><item><title><![CDATA[代码笔记：ln创建链接]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/145496156</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/145496156</guid><author>blueag1e</author><pubDate>Fri, 07 Feb 2025 15:36:25 +0800</pubDate><description><![CDATA[ln链接]]></description><category></category></item><item><title><![CDATA[代码笔记：Linux系统上解压文件]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/142772318</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/142772318</guid><author>blueag1e</author><pubDate>Thu, 17 Oct 2024 23:30:06 +0800</pubDate><description><![CDATA[【代码】代码笔记：Linux系统上解压文件。]]></description><category></category></item><item><title><![CDATA[论文阅读：AdaBins: Depth Estimation using Adaptive Bins]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/137526029</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/137526029</guid><author>blueag1e</author><pubDate>Tue, 09 Apr 2024 00:28:29 +0800</pubDate><description><![CDATA[Ddmin​dmax​)Rh×w×Cbi​∑j1N​bj′​ϵbi′​ϵ​RRcbi​d​mindmax​−dmin​bi​/2∑j1i1​bj​)d~cbi​)d∑k1N​cbk​pk​gi​logdi​−logdi​。]]></description><category></category></item><item><title><![CDATA[代码记录：3D Gaussian Splatting 渲染(Rendering)代码]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/137359858</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/137359858</guid><author>blueag1e</author><pubDate>Thu, 04 Apr 2024 01:35:33 +0800</pubDate><description><![CDATA[【代码】代码记录：3D Gaussian Splatting 渲染(Rendering)代码。]]></description><category></category></item><item><title><![CDATA[论文阅读: Visual Attention Network]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/137214822</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/137214822</guid><author>blueag1e</author><pubDate>Mon, 01 Apr 2024 03:31:16 +0800</pubDate><description><![CDATA[dK​×dK​(2d−1×2d−1)1×14H​×4W​8H​×8W​16H​×16W​32H​×32W​。]]></description><category></category></item><item><title><![CDATA[3D Gaussian Splatting for Real-Time Radiance Field Rendering(慢慢啃，还是挺复杂的)]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/136588306</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/136588306</guid><author>blueag1e</author><pubDate>Tue, 12 Mar 2024 01:32:14 +0800</pubDate><description><![CDATA[GS]]></description><category></category></item><item><title><![CDATA[论文阅读：How Do Neural Networks See Depth in Single Images?]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/136176628</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/136176628</guid><author>blueag1e</author><pubDate>Tue, 20 Feb 2024 18:50:20 +0800</pubDate><description><![CDATA[总结一下就是改变相机俯仰角和地平线位置对深度估计有影响但是影响不大。]]></description><category></category></item><item><title><![CDATA[MLP-Mixer: AN all MLP Architecture for Vision]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/136161398</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/136161398</guid><author>blueag1e</author><pubDate>Mon, 19 Feb 2024 05:51:10 +0800</pubDate><description><![CDATA[S维度就代表空间，C代表同一空间位置。]]></description><category></category></item><item><title><![CDATA[论文阅读: Semantics-guided Triplet Loss]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/135331589</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/135331589</guid><author>blueag1e</author><pubDate>Tue, 02 Jan 2024 04:01:17 +0800</pubDate><description><![CDATA[一个度量学习方法，通过浏览语义引导的局部集合去优化内在深度表示。一个新颖的特征融合模块能有效利用跨模态特异质特征。]]></description><category></category></item><item><title><![CDATA[学习使用DDP: DistributedDataParallel]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/135047342</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/135047342</guid><author>blueag1e</author><pubDate>Mon, 18 Dec 2023 02:12:58 +0800</pubDate><description><![CDATA[WORLD_SIZE:总进程数量，通常每个进程都会被分配一个唯一的 rank，而 world_size 就是所有进程的总数量。和：这两个环境变量用于指定主进程的地址和端口号。在分布式训练中，主进程通常用于协调其他进程的工作。LOCAL_RANK和LOCAL_SIZE: 这两个环境变量是 torch.distributed.launch 工具特有的，用于指定当前进程在本地节点上的 local_rank 和节点上 GPU 的总数量 local_size。RANK和WORLD_SIZE。]]></description><category></category></item><item><title><![CDATA[论文阅读：Fast-BEV: Towards Real-time On-vehicleBird’s-Eye View Perception]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/133162761</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/133162761</guid><author>blueag1e</author><pubDate>Thu, 12 Oct 2023 15:34:16 +0800</pubDate><description><![CDATA[M2BEVq∈Pxyz​kv∈F2D​uv。]]></description><category></category></item><item><title><![CDATA[论文阅读：AugGAN: Cross Domain Adaptation with GAN-based Data Augmentation]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/132984458</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/132984458</guid><author>blueag1e</author><pubDate>Fri, 22 Sep 2023 10:45:36 +0800</pubDate><description><![CDATA[[1] Huang S W, Lin C T, Chen S P, et al. Auggan: Cross domain adaptation with gan-based data augmentation[C]//Proceedings of the European Conference on Computer Vision (ECCV). 2018: 718-731.]]></description><category></category></item><item><title><![CDATA[论文笔记：Reinforcing Local Structure Perception for Monocular Depth Estimation]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/131811112</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/131811112</guid><author>blueag1e</author><pubDate>Thu, 07 Sep 2023 11:15:30 +0800</pubDate><description><![CDATA[在m个不同尺度上采样三个点。]]></description><category></category></item><item><title><![CDATA[论文阅读：SERE: Exploring Feature Self-relation for Self-supervised Transformer]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/132671702</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/132671702</guid><author>blueag1e</author><pubDate>Tue, 05 Sep 2023 22:29:06 +0800</pubDate><description><![CDATA[使用自监督学习为卷积网络（CNN）学习表示已经被验证对视觉任务有效。作为CNN的一种替代方案，视觉变换器（ViT）具有强大的表示能力，具有空间自注意力和通道级前馈网络。最近的研究表明，自监督学习有助于释放ViT的巨大潜力。然而，大多数研究仍然遵循为CNN设计的自监督策略，例如实例级别的样本区分，但它们忽视了ViT的特性。我们观察到，关于空间和通道维度的关系建模使ViT与其他网络有所不同。为了强化这一特性，我们探索了特征自关系（SElf-RElation，SERE）来训练自监督ViT。]]></description><category></category></item><item><title><![CDATA[论文阅读：Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/132508526</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/132508526</guid><author>blueag1e</author><pubDate>Sat, 26 Aug 2023 11:30:36 +0800</pubDate><description><![CDATA[我们定义一个模型fff, 输入xxx输出aaa。-定义每一个Task -TTT包含一个损失函数LLL, 一个原始观察qx1q(x_1)qx1​, 一个状态转移分布qx1∣xtatqx1​∣xt​at​以及集长度HHH。在监督任务中H=1(也就是说当前的a只和当前的x有关)。在我们的元学习场景中，我们考虑了一个跨任务的分布pTp(T)pT，我们希望我们的模型能够适应这个分布。]]></description><category></category></item><item><title><![CDATA[论文笔记Monocular Depth Estimation through Virtual-world Supervision and Real-world SfM Self-Supervision]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/131661711</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/131661711</guid><author>blueag1e</author><pubDate>Tue, 11 Jul 2023 20:22:14 +0800</pubDate><description><![CDATA[Ψθ;x→dθ∗minθ​Lθ;XrXsYs)Xrxtr​t1Nr​NrXsxts​t1Ns​Nr​Ψθ;x)Ψθ;x)θencθpyrθdecLLsfθVsf;XrLspθXs;YsLDAθencVDA;XrXs)LsfθVsf;Xr)LspθXs;Ys)dts​p≥dma。]]></description><category></category></item><item><title><![CDATA[论文阅读：Dense Depth Priors for Neural Radiance Fields from Sparse Input Views]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/131294928</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/131294928</guid><author>blueag1e</author><pubDate>Mon, 19 Jun 2023 22:16:35 +0800</pubDate><description><![CDATA[CVPR2022。]]></description><category></category></item><item><title><![CDATA[论文阅读：Scale-aware and High Quality Monocular Depth Estimation Without Real World Labels]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/131225245</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/131225245</guid><author>blueag1e</author><pubDate>Thu, 15 Jun 2023 14:58:12 +0800</pubDate><description><![CDATA[CVPR2022]]></description><category></category></item><item><title><![CDATA[论文笔记：Normalizing Flows for Probabilistic Modeling and Inference]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/131042217</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/131042217</guid><author>blueag1e</author><pubDate>Mon, 05 Jun 2023 11:41:55 +0800</pubDate><description><![CDATA[Normalizing Flow 笔记]]></description><category></category></item><item><title><![CDATA[【代码笔记】准备Kitti raw中的Pose数据]]></title><link>https://blog.youkuaiyun.com/blueag1e/article/details/130842465</link><guid>https://blog.youkuaiyun.com/blueag1e/article/details/130842465</guid><author>blueag1e</author><pubDate>Wed, 24 May 2023 11:02:43 +0800</pubDate><description><![CDATA[将Kitti_raw转化为python可用数据]]></description><category></category></item></channel></rss>