小花生的博客

验证预处理一致性import osimport cv2import numpy as npimport PILimport torchfrom PIL import Imageimport torchvision.transforms as transformsclass RGBToBGR(): def __call__(self, im): assert im.mode == 'RGB' r, g, b = [im.getchannel(i)

PyTorch学习（10）:训练技巧Pytorch官方文档： https://pytorch-cn.readthedocs.io/zh/latest/参考：  https://efficientdl.com/faster-deep-learning-in-pytorch-a-guide/  https://www.reddit.com/r/MachineLearning/comments/n9fti7/d_a_few_helpful_pytorch_tips  _examples_included

PyTorch学习（9）:实战Pytorch官方文档： https://pytorch-cn.readthedocs.io/zh/latest/Pytorch学习文档： https://github.com/tensor-yu/PyTorch_TutorialPytorch模型库： https://github.com/pytorch/vision/tree/master/torchvision/models文章目录PyTorch学习（9）:实战前言1.ShuffleNet系列（1）数据处理（2

PyTorch学习（8）:模型保存和加载Pytorch官方文档： https://pytorch-cn.readthedocs.io/zh/latest/Pytorch学习文档： https://github.com/tensor-yu/PyTorch_Tutorial参考： https://mp.weixin.qq.com/s/iYuMcbnJqVUMDMvwDEPs6Q    https://www.dtmao.cc/news_show_329038.shtml文章目录PyTorch学习（

PyTorch学习（7）:学习率Pytorch官方文档： https://pytorch-cn.readthedocs.io/zh/latest/Pytorch学习文档： https://github.com/tensor-yu/PyTorch_Tutorial参考： https://pytorch.org/docs/stable/optim.html?highlight=lr_scheduler    https://blog.youkuaiyun.com/u011995719/article/detail

PyTorch学习（6）:优化算法Pytorch官方文档： https://pytorch-cn.readthedocs.io/zh/latest/Pytorch学习文档： https://github.com/tensor-yu/PyTorch_Tutorial参考： https://pytorch-cn.readthedocs.io/zh/latest/package_references/torch-optim/#optimizer    https://zhuanlan.zhihu.com

PyTorch学习（5）:损失函数Pytorch官方文档： https://pytorch-cn.readthedocs.io/zh/latest/Pytorch学习文档： https://github.com/tensor-yu/PyTorch_Tutorial参考： https://pytorch.org/docs/stable/nn.html    https://pytorch-cn.readthedocs.io/zh/latest/package_references/torch-nn/

PyTorch学习（4）:模型Pytorch官方文档： https://pytorch-cn.readthedocs.io/zh/latest/Pytorch学习文档： https://github.com/tensor-yu/PyTorch_TutorialPytorch模型库： https://github.com/pytorch/vision/tree/master/torchvision/models参考： https://zhuanlan.zhihu.com/p/27401516 

PyTorch学习（3）:数据预处理Pytorch官方文档： https://pytorch-cn.readthedocs.io/zh/latest/Pytorch学习文档： https://github.com/tensor-yu/PyTorch_Tutorial参考： https://blog.youkuaiyun.com/u011995719/article/details/85103712   https://www.cnblogs.com/mengfu188/p/13561693.html   h

PyTorch学习（2）:数据加载机制Pytorch官方文档：https://pytorch-cn.readthedocs.io/zh/latest/Pytorch学习文档：https://github.com/tensor-yu/PyTorch_Tutorial参考：https://blog.youkuaiyun.com/u011995719/article/details/85102770文章目录PyTorch学习（2）:数据加载机制前言1.Dataset类2.构建自定义Dataset子类3.DataL

PyTorch学习（1）：基础知识Pytorch官方文档： https://pytorch-cn.readthedocs.io/zh/latest/Pytorch学习文档： https://github.com/tensor-yu/PyTorch_Tutorial参考： https://blog.youkuaiyun.com/u011995719/article/details/85102770   https://blog.youkuaiyun.com/weicao1990/article/details/93177

运行torch模型时，报错信息如下：OMP: Error #15: Initializing libiomp5md.dll, but found libiomp5md.dll already initialized.OMP: Hint This means that multiple copies of the OpenMP runtime have been linked into the program. That is dangerous, since it can degrade perform

运行出现错误：RuntimeError: CUDA out of memory. Tried to allocate 500.00 MiB (GPU 0; 4.00 GiB total capacity; 204.39 MiB already allocated; 363.42 MiB free; 373.61 MiB cached)解决方法： 释放显存torch.cuda.empty_cache()如果还是解决不了：（1）增加显存（2）切换为CPU运行...

Github下载的Pytorch0.4.1的代码，linux运行没问题，转到windows出现错误！File “D:\Anaconda3.7\lib\multiprocessing\process.py”, line 112, in startself._popen = self._Popen(self)File “D:\Anaconda3.7\lib\multiprocessing\context.py”, line 223, in _Popenreturn _default_context.g

训练PyTorch代码出现：Cannot allocate memory 错误。解决方法： 修改多线程数量，设置num_workers，系统默认的数量是4，改成1之后，没有效果，后面改成0，问题解决。...

PyTorch安装按照官网上的命令安装即可。最新版本：https://pytorch.org/get-started/locally/旧版本：https://pytorch.org/get-started/previous-versions/报错的画，参考：https://blog.youkuaiyun.com/songchunxiao1991/article/details/95192063...

Pytorch - TORCH.NN.INIT 参数初始化方法路径：https://pytorch.org/docs/master/nn.init.html#nn-init-doc初始化函数：torch.nn.init# -*- coding: utf-8 -*-"""Created on 2019@author: fancp"""import torch import...

Pytorch版本代码修正概述：老版本代码（0.3.0）修正为（0.4.0）1.UserWarning: nn.init.xavier_uniform is now deprecated in favor of nn.init.xavier_uniform_.init.xavier_uniform(param)注：init.xavier_uniform 改为 init.xavier_uni...

代码：# -*- coding: utf-8 -*-import heapqnums = [1, 8, 2, 23, 7, -4, 18, 23, 24, 37, 2]# 最大的3个数的索引max_num_index_list = map(nums.index, heapq.nlargest(3, nums))# 最小的3个数的索引min_num_index_list = ma...

代码起始处添加：即可import matplotlib.pyplot as pltimport seaborn as snsfrom matplotlib.font_manager import FontPropertiesmyfont=FontProperties(fname=r’C:\Windows\Fonts\simhei.ttf’,size=15)sns.set(font=myf...

(base) C:\Users\pxsj_admin>pip install torchvisionCollecting torchvisionUsing cached https://files.pythonhosted.org/packages/fb/01/03fd7e503c16b3dc262483e5555ad40974ab5da8b9879e164b56c1f4ef6f/tor...