- 博客(293)
- 资源 (11)
- 收藏
- 关注
原创 ubuntu系统文件误删(/lib/x86_64-linux-gnu/libc.so.6)修复方案 [成功解决]
Linux系统修复:libc.so.6缺失解决方案当Linux系统出现libc.so.6缺失导致命令无法使用时,可通过以下步骤修复:使用对应版本Ubuntu启动盘进入Try Ubuntu模式挂载原系统分区:sudo mount /dev/sda1 UDriveSys从启动盘复制libc.so.6到原系统:sudo cp /lib/x86_64-linux-gnu/libc.so.6 /挂载路径/lib/x86_64-linux-gnu同时复制/sbin/init文件以防引导问题重启系统后可能
2025-06-05 15:36:49
586
原创 /usr/lib/x86_64-linux-gnu/libstdc++.so.6: version`GLIBCXX_3.4.29‘ not found [成功解决] (错误方法易至系统故障)
TensorFlow运行时出现GLIBCXX_3.4.29缺失报错,原因是系统GLIBC版本过低。有效解决方法是添加Ubuntu jammy源后升级libc6:先修改/etc/apt/sources.list添加阿里云源,执行更新安装后即可解决。文章警告勿采用复制高版本libstdc++.so.6的错误方案,该操作会导致系统兼容性问题(如GLIBC_2.34缺失),严重时需重装系统。最后提示不要删除/lib/x86_64-linux-gnu/libc.so.6文件。
2025-06-05 04:23:48
342
原创 PIL 报错AttributeError: module ‘PIL.Image‘ has no attribute ‘ANTIALIAS‘ 解决方案
PIL 报错 ANTIALIAS
2024-11-29 00:58:57
751
原创 Pose Estimation问题汇总:albumentations,MultiScaleDeformableAttention,crowdposetools,GLIBCXX_3.4.29,sm_35
pose estimation环境配置问题汇总,:albumentations, MultiScaleDeformableAttention,crowdposetools,GLIBCXX_3.4.29,mmcv 安装 no
2024-08-13 20:00:05
102
原创 卷积报错:AttributeError: ‘Conv2d‘ object has no attribute ‘total_ops‘ (已解)
卷积报错:AttributeError: 'Conv2d' object has no attribute 'total_ops', Only Tensors created explicitly by the user (graph leaves) support the deepcopy protocol at the moment
2024-05-21 14:42:40
1398
1
原创 CUDA error: the provided PTX was compiled with an unsupported toolchain.(已解决:多机多卡并行训练配置问题)
CUDA error: the provided PTX was compiled with an unsupported toolchain. 可行解决
2024-05-14 14:56:10
1401
1
原创 [ WACV24]SBCFormer:快速、轻量级 神经网络模型 Efficient CNN-ViT hybrid neural network backbone
快速、轻量级hybrid神经网络for边缘AI,Vision Transformer
2023-11-08 16:35:13
737
6
原创 IEEE PDF eXpress报错: TimesNewRomanPS-ItalicMT, TimesNewRomanPSMT is not embedded; Font Helvetica
IEEE PDF eXpress报错:TimesNewRomanPS-ItalicMT, TimesNewRomanPSMT, Font Helvetica are not embedded
2023-11-05 18:26:49
1395
2
原创 pytorch底层矩阵运算:How to check if PyTorch or Numpy is using OpenBLAS or MKL? &线性代数库BLAS(cblas_gemm示例)
检测pytorch使用哪种线性代数运算库(矩阵或矢量间线性代数); BLAS, MKL, OpenBLAS及cblas_gemm矩阵运算代码.
2023-08-14 21:50:25
633
原创 分布式训练Warning: Grad strides do not match bucket view strides. This may indicate grad was not
由而导致的Grad strides不匹配问题。
2023-08-08 01:46:20
2928
原创 pytorch报错 ERROR:torch.distributed.elastic.multiprocessing.api:failed (exitcode: 1) local_rank:....
pytorch报错 ERROR:torch.distributed.elastic.multiprocessing.api:failed (exitcode: 1) local_rank:....
2023-06-26 14:08:28
13875
原创 多GPU训练报错:at least two devices cuda:1 and cuda:0;Caught ValueError in replica 0 on device 0
DataParallel和DistributedDataParallel多GPU训练时报错,RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:1 and cuda:0!
2023-04-26 13:06:39
1728
原创 分布式训练模型保存torch.save()、加载model.load_state_dict问题解析:Missing key(s) in state_dict,Unexpected key()
分布式训练模型保存torch.save()、加载model.load_state_dict常见问题:Missing key(s) in state_dict,Unexpected key()
2022-11-18 15:06:12
1225
原创 FLOPS(每秒浮点运算次数), TFLOP,Statistical vs. Computational Efficiency
FLOPS全称是floating-point operation per second,每秒浮点运算次数,用于衡量计算机的算力和执行效能。The difference between pure computational efficiency and statistical efficiency.
2022-10-20 03:37:32
5661
原创 Ubuntu 上VS Code更新后 Remote-SSH 无法连接服务器:failed to install the vs code server!
正常使用VS code SSH,退出VS code重新启动后,出现错误提示:Can't connect to the server,failed to install the vs code server.
2022-10-17 12:35:47
2769
原创 报Warning: variables__flops__or_params__ are already defined for the---ptflops can affect your code!
Warning: variables__flops__or_params__ are already defined for the---ptflops can affect your code!
2022-10-12 23:22:38
721
原创 .sh脚本调用服务器上conda虚拟环境 & CUDA版本与服务器GPU driver不匹配(ubuntu系统、pytorch)
无法调用服务器上的虚拟环境pytorch, CUDA initialization: The NVIDIA driver on your system is too old (found version 10010)
2022-10-04 00:27:27
851
原创 卷积复杂度计算、im2col算法、Normalization【实现方法解析】
计算复杂度分析,卷积官方代码分析,Batch Norm, Layer Norm, Group Norm 理解
2022-09-15 16:46:27
1152
原创 pytorch报错:RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)`
多GPU分布式运行pytorch程序报错:RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling ‘cublasCreate(handle)`
2022-07-13 16:22:27
7097
1
原创 PyTorch 分布式训练 (DP/DDP/torchrun/多机多卡) <笔记总结>
PyTorch 分布式训练问题总结: (DP/DDP/torchrun/多机多卡)
2022-06-29 21:21:55
3793
原创 Python代码 时间消耗latency和空间(memory)消耗(详细剖析、求解)
Python代码 时间消耗和空间(内存)消耗(详细剖析、求解)Memit lineprofiler
2022-05-18 10:19:48
378
1
原创 arm-linux-gnueabihf-gcc 安装 OpenBLAS
参考:1.MIPS(arm)交叉编译 OpenBLAS 编译和安装 2.ubuntu18.04 安装OpenBLAS问题一:LAPACK make fails: “recipe for target ‘znep.out‘ failed” errorgfortran -O2 -frecursive -c -o zunt03.o zunt03.fgfortran -o xeigtstz zchkee.o zbdt01.o zbdt02.o zbdt03.o zbdt0...
2022-01-19 18:06:32
2079
原创 pytorch C++ 扩展:torch.utils.cpp_extension
torch.utils.cpp_extension-PyTorch 1.0 中文文档 & 教程
2021-12-24 14:39:46
870
原创 ARM NEON 程序debug: error识别arm neon intrinsics 命令, LD_LIBRARY_PATH shouldn‘t
/usr/lib/gcc/arm-linux-gnueabihf/6/include/arm_neon.h:5792:1: error: inlining failed in call to always_inline ‘vdupq_n_s32’: target specific option mismatch vdupq_n_s32 (int32_t __a) ^~~~~~~~~~~neon_test.c:94:13: note: called from here int32x4_t sum.
2021-12-24 10:50:46
1526
原创 ARM Neon 编程笔记一(ARM NEON Intrinsics, SIMD运算, 优化心得)
1. ARM Neon Intrinsics 学习指北:从入门、进阶到学个通透1.入门:基本能上手写Intrinsics1.1 Neon介绍、简明案例与编程惯例1.2 如何检索Intrinsics1.3 优化效果案例1.4 如何在Android应用Neon2. 进阶:注意细节处理,学习常用算子的实现2.1 与Neon相关的ARM体系结构2.2 对非整数倍元素个数(leftovers)的处理技巧2.3 算子源码学习(ncnn库,AI方向)2.4 算子源码学习(Nvidia caro..
2021-12-07 15:03:27
1876
2
原创 树莓派 切换python版本、安装pytorch、vscode、opencv-python、matplotlib
参考:1、将树莓派中默认python2切换python3_浪迹天涯的博客-CSDN博客 2、树莓派:树莓派4B+安装Pytorch版本1.6(python3.7) 3、树莓派4b安装vscode亲测可行。--相关操作(需要时候再转换):把树莓派4B系统修改为64位...
2021-10-12 00:34:44
2311
原创 基于python numpy 手写实现 Batch Normalization
二 、基于python numpy,下面我们自己实现批量归一化层。import timeimport torchfrom torch import nn, optimimport torch.nn.functional as Fimport syssys.path.append("..") import d2lzh_pytorch as d2ldevice = torch.device('cuda' if torch.cuda.is_available() else ...
2021-10-04 20:29:27
1149
原创 transformer和functional.py错:Argument interpolation should be of type InterpolationMode instead of int
transformer和functional.py报错:Argument interpolation should be of type InterpolationMode instead of int
2021-09-24 23:50:19
5921
3
原创 Pytorch分布式报错1.non-zero exit status,2.cuDNN error:CUDNN_STATUS_INTERNAL_,3.CUDA error:illegal memory
1.returned non-zero exit status 1.One epoch之后报错,信息如下:RuntimeError: Expected to have finished reduction in the prior iteration before starting a new one. This error indicates that your module has parameters that were not used in producing loss. You..
2021-07-26 15:53:54
2881
转载 Rethinking Semantic Segmentation from a Sequence-to-Sequence Perspectivewith Transformers
重新思考语义分割范式——SETR转载from:https://zhuanlan.zhihu.com/p/348418189 请多支持原创一、论文信息标题:《Rethinking Semantic Segmentation from a Sequence-to-Sequence Perspectivewith Transformers》作者:Sixiao Zheng et al.(复旦大学 & 牛津大学 & 萨里大学 & 腾讯优图 & 脸书)文章:h...
2021-07-04 10:50:45
1352
原创 Ubuntu18.04 长按Back Space键 只能删除一个字符
问题描述:Ubuntu 系统一直使用正常,突然键盘只能按一个字符一个字符的删除,键盘迟钝。(可能是系统更新造成的系统设置重置)解决方案如图: 确保2处Repeat Keys 打开, 以及 Typing Assist 关闭。问题即可解决。...
2021-06-03 19:01:43
1331
Matlab 2015b Win64 Crack ,Matlab2015b 破解文档,已破解成功
2016-06-03
Large_Margin_Nearest_Neighbor_Classifiers
2013-12-09
空空如也
TA创建的收藏夹 TA关注的收藏夹
TA关注的人