BlueagleAI-优快云博客

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。

2023-12-18 02:12:58 1659

原创论文阅读：Fast-BEV: Towards Real-time On-vehicleBird’s-Eye View Perception

M2BEVq∈Pxyzkv∈F2Duv。

2023-10-12 15:34:16 296

原创论文阅读：AugGAN: Cross Domain Adaptation with GAN-based Data Augmentation

[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.

2023-09-22 10:45:36 581 1

原创论文笔记：Reinforcing Local Structure Perception for Monocular Depth Estimation

在m个不同尺度上采样三个点。

2023-09-07 11:15:30 232 1

原创论文阅读：SERE: Exploring Feature Self-relation for Self-supervised Transformer

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。

2023-09-05 22:29:06 661

原创论文阅读：Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks

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∣xtat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。

2023-08-26 11:30:36 1265

原创论文笔记Monocular Depth Estimation through Virtual-world Supervision and Real-world SfM Self-Supervision

Ψθ;x→dθ∗minθLθ;XrXsYs)Xrxtrt1NrNrXsxtst1NsNrΨθ;x)Ψθ;x)θencθpyrθdecLLsfθVsf;XrLspθXs;YsLDAθencVDA;XrXs)LsfθVsf;Xr)LspθXs;Ys)dtsp≥dma。

2023-07-11 20:22:14 180 2

原创论文阅读：Dense Depth Priors for Neural Radiance Fields from Sparse Input Views

CVPR2022。

2023-06-19 22:16:35 2109 2

原创论文阅读：Scale-aware and High Quality Monocular Depth Estimation Without Real World Labels

CVPR2022

2023-06-15 14:58:12 110 1

原创论文笔记：Normalizing Flows for Probabilistic Modeling and Inference

Normalizing Flow 笔记

2023-06-05 11:41:55 1546 2

翻译【代码笔记】准备Kitti raw中的Pose数据

将Kitti_raw转化为python可用数据

2023-05-24 11:02:43 586

原创论文阅读: Redesign Self-Supervised MonoDepth for Full-Scale Depth Prediction for Autonomous Driving

对FS-Dpeth的记录

2023-05-23 21:38:19 156 1

原创论文阅读: NeRF−−: Neural Radiance Fields Without Known Camera Parameters

NeRF−−: Neural Radiance Fields Without Known Camera Parameters

2023-03-07 23:17:27 572 1

原创论文阅读：Direct Voxel Grid Optimization: Super-fast Convergence for Radiance Fields Reconstruction

Direct Voxel Grid Optimization: Super-fast Convergence for Radiance Fields Reconstruction

2023-03-04 18:32:12 1083 1

原创论文阅读：MPViT : Multi-Path Vision Transformer for Dense Prediction

MPViT : Multi-Path Vision Transformer for Dense Prediction

2023-03-01 11:08:36 923 1

原创论文阅读：Self-Supervised Monocular Depth Estimation with Internal Feature Fusion(DIFFNet)

Self-Supervised Monocular Depth Estimation with Internal Feature Fusion

2023-02-28 16:36:39 811 1

原创论文阅读：NeRF++: ANALYZING AND IMPROVING NEURAL RADIANCE FIELDS

NeRF++: ANALYZING AND IMPROVING NEURAL RADIANCE FIELDS

2023-02-27 11:22:28 660 1

原创论文阅读：RA-Depth: Resolution Adaptive Self-Supervised Monocular Depth Estimation

RA-Depth: Resolution Adaptive Self-Supervised Monocular Depth Estimation

2023-02-22 14:33:45 700 4

原创论文阅读：pixelNeRF: Neural Radiance Fields from One or Few Images

pixelNeRF: Neural Radiance Fields from One or Few Images

2023-02-20 22:17:07 584 1

原创论文阅读：InfoNeRF: Ray Entropy Minimization for Few-Shot Neural Volume Rendering

InfoNeRF: Ray Entropy Minimization for Few-Shot Neural Volume Rendering

2023-02-20 15:43:36 944 1

原创论文笔记：RegNeRF: Regularizing Neural Radiance Fields for View Synthesis from Sparse Inputs

RegNeRF: Regularizing Neural Radiance Fields for View Synthesis from Sparse Inputs

2023-02-17 22:13:30 581 2

原创论文笔记：NeuLF: Efficient Novel View Synthesis with Neural 4D Light Fight

NeuLF: Efficient Novel View Synthesis with Neural 4D Light Fight

2023-02-15 21:12:42 552 1

原创论文笔记：Vision Transformers for Dense Prediction

Vision Transformers for Dense Prediction

2023-02-15 10:50:06 946 1

原创论文笔记： Monocular Depth Estimation: a Review of the 2022 State of the Art

Monocular Depth Estimation: a Review of the 2022 State of the Art

2023-02-14 13:20:21 1129 4

原创论文阅读：GeoAug: Data Augmentation for Few-Shot NeRF with Geometry Constrain

GeoAug: Data Augmentation for Few-Shot NeRF with Geometry Constrain

2023-02-13 21:57:45 521 1

原创论文阅读：MINE: Towards Continuous Depth MPI with NeRF for Novel View Synthes

MINE: Towards Continuous Depth MPI with NeRF for Novel View Synthes

2023-02-11 10:40:01 790 1

原创论文笔记：NerfingMVS: Guided Optimization of Neural Radiance Fields for Indoor Multi-view Stereo

NerfingMVS: Guided Optimization of Neural Radiance Fields for Indoor Multi-view Stereo

2023-02-09 15:30:13 398 1

原创论文笔记：Depth-supervised NeRF: Fewer Views and Faster Training for Free

Depth-supervised NeRF: Fewer Views and Faster Training for Free

2023-02-08 21:11:34 1787 1

原创论文笔记：D-NeRF:Neural Radiance Fields for Dynamic Scenes

D-NeRF:Neural Radiance Fields for Dynamic Scenes

2023-02-07 23:04:36 1605 4

原创论文笔记：HR-Depth:High Resolution Self-Supervised Monocular Depth Estimation

HR-Depth:High Resolution Self-Supervised Monocular Depth Estimation

2023-02-07 14:46:28 1250 1

原创论文笔记：Excavating the Potential Capacity of Self-Supervised Monocular Depth Estimation

Excavating the Potential Capacity of Self-Supervised Monocular Depth Estimation

2023-02-07 10:39:30 371 1

空空如也

空空如也