byobu

byobu/tmux是screen的增强版, 敲入byobu命令直接创建和恢复会话，其他快捷键可与screen兼容

安装byobu:

sudo apt install byobu

新建一个byobu会话：

byobu new -s

关闭一个byobu会话：

byobu kill-sesssion -t

恢复某个byobu会话：

byobu a -t

查看有哪些byobu会话：

byobu ls

常用快捷键

shortcut	usage
F2	Create a new window 打开一个新的窗口
F3	Move to the previous window 进入前一个窗口
F4	Move to the next window 进入后一个窗口
F5	Refresh all status notifications
F6	Detach from the session and logout 断开链接
F7	Enter copy/scrollback mode 进入scrollback模式
F8	Rename the current window
F9	Launch the Byobu Configuration Menu
F12	Lock this terminal
Alt-Pageup	Scroll back through this window’s history
Alt-Pagedown	Scroll forward through this window’s history
shift - F1	帮助
Shift-F2	Split the screen horizontally 横切割新建个窗口
Shift-F3	Move focus to the next split
Shift-F4	Move focus to the previous split
Shift-F6	Detach from the session, but do not logout
Shift-F5	Collapse all splits
Shift+F11	最大化其中一个窗口
Shift-F12	Toggle all of Byobu’s keybindings on or off
Ctrl+F2	Split the screen vertically 竖切割新建个窗口
Ctrl+F5	Reconnect any SSH/GPG sockets or agents
Ctrl+F6	Kill window

F7 scrollback模式下操作

I hit F7 to enter scrollback mode,
Space to start selecting,
g to scroll to the top of the buffer (thanks @GeorgeMarian)
Enter to copy (to byobu’s clipboard, not a terminal/system one),
then cat > my-byobu-dump.txt in the terminal,
Alt+Insert or ctrl+A,] to paste (again, from byobu’s clipboard)
Ctrl+D to close the file.

refer

# !/bin/bash
# Usage: force_cmd.sh pull/push dir max_tries
cd $2
if [ ! $3 ]
thenmax=10
elsemax=$3
fi
echo "max=$max"num=0
#git $1# error code 128: unable to access
while [[ $num -le $max ]]
doerror=0if ! git $1; then error=1; finum=$[$num+1]echo $numif [ $error = 0 ]thenexitfi
done

# !/bin/bash
# Usage:
#     alias fgit='bash ...force_git.sh'
# fgit [max_tries] args...
if ! [[ $1 =~ ^-?[0-9]+$ ]]
thenmax=10cmd=${@:1}
elsemax=$1cmd=${@:2}
fi
echo "max=$max"
echo "cmd=git $cmd"
num=0# error code 128: unable to access
while [[ $num -le $max ]]
doerror=0if ! git $cmd; then error=1; finum=$[$num+1]echo $numif [ $error = 0 ]thenexitfi
done

https://oldpan.me/archives/pytorch-to-use-multiple-gpus
https://zhuanlan.zhihu.com/p/86441879
https://zhuanlan.zhihu.com/p/234293510
https://zhuanlan.zhihu.com/p/145427849level set , water shed , point cloud, convex optim , unsupervised , deform

#  Copyright (c) 2020. The Medical Image Computing (MIC) Lab, 陶豪毅
#
#  Licensed under the Apache License, Version 2.0 (the "License");
#  you may not use this file except in compliance with the License.
#  You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
#  Unless required by applicable law or agreed to in writing, software
#  distributed under the License is distributed on an "AS IS" BASIS,
#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#  See the License for the specific language governing permissions and
#  limitations under the License.from typing import Union, List
import torch
import torch.nn as nnfrom medtk.model.nnModules import ComponentModule, BlockModule
from medtk.model.nd import MaxPoolNd
from medtk.model.blocks import ConvNormAct, VResConvNormAct, \BasicBlockNd, BottleneckNd, \SEBasicBlockNd, SEBottleneckNdclass ConvLayer(BlockModule):def __init__(self, dim, in_channels, out_channels, stride, num_blocks, groups=1, base_width=64, dilation=1):super().__init__()self.blocks = []for i in range(num_blocks):if i == 0:self.blocks.append(ConvNormAct(dim, in_channels, out_channels, kernel_size=3, padding=1, stride=stride))else:self.blocks.append(ConvNormAct(dim, out_channels, out_channels, kernel_size=3, padding=1))self.blocks.append(ConvNormAct(dim, out_channels, out_channels, kernel_size=3, padding=1))for i, m in enumerate(self.blocks):self.add_module(str(i), m)def forward(self, x):for layer in self.blocks:x = layer(x)return xclass VResConvLayer(BlockModule):def __init__(self, dim, in_channels, out_channels, stride, num_blocks, groups=1, base_width=64, dilation=1):super().__init__()self.blocks = []for i in range(num_blocks):if i == 0:self.blocks.append(ConvNormAct(dim, in_channels, out_channels, kernel_size=3, padding=1, stride=stride))else:self.blocks.append(VResConvNormAct(dim, out_channels, out_channels, kernel_size=3, padding=1))for i, m in enumerate(self.blocks):self.add_module(str(i), m)def forward(self, x):for layer in self.blocks:x = layer(x)return xclass ResidualLayer(BlockModule):BLOCK = BasicBlockNddef __init__(self, dim, in_channels, out_channels, stride, num_blocks, groups=1, base_width=64, dilation=1):super(ResidualLayer, self).__init__()downsample = Noneif stride != 1 or in_channels != out_channels * self.BLOCK.expansion:downsample = nn.Sequential(self.build_conv(dim, in_channels, out_channels * self.BLOCK.expansion,kernel_size=1, stride=stride, bias=False),self.build_norm(dim, out_channels * self.BLOCK.expansion),)self.blocks = nn.ModuleList([self.BLOCK(dim,in_planes=in_channels,planes=out_channels,stride=stride,dilation=1,downsample=downsample,groups=1,width_per_group=64)])in_planes = out_channels * self.BLOCK.expansionfor i in range(1, num_blocks):self.blocks.append(self.BLOCK(dim,in_planes=in_planes,planes=out_channels,stride=1,dilation=1,groups=1,width_per_group=64))for i, m in enumerate(self.blocks):self.add_module(str(i), m)def forward(self, x):for i in range(len(self.blocks)):layer = getattr(self, str(i))x = layer(x)return xclass ResidualBottleneckLayer(ResidualLayer):BLOCK = BottleneckNddef __init__(self, dim, in_channels, out_channels, stride, num_blocks, groups=1, base_width=64, dilation=1):out_channels = out_channels // self.BLOCK.expansionsuper(ResidualBottleneckLayer, self).__init__(dim, in_channels, out_channels,stride, num_blocks, groups, base_width, dilation)class SEResidualLayer(ResidualLayer):BLOCK = SEBasicBlockNddef __init__(self, dim, in_channels, out_channels, stride, num_blocks, groups=1, base_width=64, dilation=1):super(SEResidualLayer, self).__init__(dim, in_channels, out_channels,stride, num_blocks, groups, base_width, dilation)class SEResidualBottleneckLayer(ResidualLayer):BLOCK = SEBottleneckNddef __init__(self, dim, in_channels, out_channels, stride, num_blocks, groups=1, base_width=64, dilation=1):out_channels = out_channels // self.BLOCK.expansionsuper(SEResidualBottleneckLayer, self).__init__(dim, in_channels, out_channels,stride, num_blocks, groups, base_width, dilation)class Encoder(ComponentModule):"""support list:- Vanilla UNet- ResNet- ResNeXt-"""LAYERS = {'conv': (ConvLayer, 1),  # UNet, VNet'v_conv': (VResConvLayer, 1),  # VBNet'res': (ResidualLayer, 1),  # ResNet 18, 34'b_res': (ResidualBottleneckLayer, 4),  # ResNet or ResNeXt ge than 50'se_res': (SEResidualLayer, 1),  # SEResNet 18, 34'se_b_res': (SEResidualBottleneckLayer, 4),}def __init__(self,dim: int,in_channels: int,features=(16, 32, 64, 128),strides=(1, 2, 2, 2),dilations=(1, 1, 1, 1),num_blocks=(1, 1, 1, 1),out_indices=(0, 1, 2, 3),layer_type='conv',groups=1,width_per_group=64,first_conv=(64, 7, 1),downsample=False):super(Encoder, self).__init__()assert isinstance(out_indices, (list, tuple)), \'out_indices must be a list/tuple but get a {}'.format(type(out_indices))assert max(out_indices) < len(strides), "max out_index must smaller than stages"assert len(strides) == len(num_blocks) == len(features)assert layer_type in self.LAYERS.keys()self.first_features, self.first_kernel, self.first_stride = first_convself.downsample = downsampleself.dim = dimself.in_channels = in_channelsself.features = featuresself.strides = stridesself.dilations = dilationsself.num_blocks = num_blocksself.out_indices = out_indicesself.stages = len(self.strides)self.groups = groupsself.width_per_group = width_per_groupself.layer_type = layer_typeself.layer, self.expansion = self.LAYERS[layer_type]self.conv1 = self.build_conv(dim, self.in_channels,self.first_features,kernel_size=self.first_kernel,stride=self.first_stride,padding=self.first_kernel // 2,bias='res' not in self.layer_type)self.bn1 = self.build_norm(self.dim, self.first_features)self.relu = self.build_act()self.maxpool = MaxPoolNd(self.dim)(kernel_size=3, stride=2, padding=1)self.layers = self.init_layers()# self.init_weights()def init_layers(self):layers = nn.ModuleList()in_planes = self.first_featuresfor i in range(self.stages):layer_name = 'layer{}'.format(i + 1)layer = self.layer(self.dim,in_planes,self.features[i],stride=self.strides[i],num_blocks=self.num_blocks[i],groups=self.groups,base_width=self.width_per_group,dilation=self.dilations[i])in_planes = self.features[i]self.add_module(layer_name, layer)layers.append(layer)return layersdef init_weights(self):for m in self.modules():if self.is_conv(self.dim, m):nn.init.kaiming_normal_(m.weight, 1e-2)if m.bias is not None:nn.init.constant_(m.bias, 0)elif self.is_norm(self.dim, m):nn.init.normal_(m.weight, 1.0, 0.02)m.bias.data.zero_()def forward(self, inputs):x = self.conv1(inputs)x = self.bn1(x)x = self.relu(x)if self.downsample:x = self.maxpool(x)outs = []for i in range(self.stages):layer_name = 'layer{}'.format(i + 1)layer = getattr(self, layer_name)x = layer(x)if i in self.out_indices:outs.append(x)return outsif __name__ == "__main__":def init_seed(SEED):torch.manual_seed(SEED)torch.cuda.manual_seed_all(SEED)torch.backends.cudnn.deterministic = Truetorch.backends.cudnn.benchmark = Falseinit_seed(666)# UNet = Encoder(#     dim=2,#     in_channels=3,#     features=(32, 64, 128, 256),#     strides=(1, 2, 2, 2),#     dilations=(1, 1, 1, 1),#     num_blocks=(1, 1, 1, 1),#     out_indices=(0, 1, 2, 3),#     layer_type='conv'# )# model = UNet# VNet = Encoder(#     dim=2,#     in_channels=3,#     features=(32, 64, 128, 256),#     strides=(1, 2, 2, 2),#     dilations=(1, 1, 1, 1),#     num_blocks=(1, 2, 3, 4),#     out_indices=(0, 1, 2, 3),#     layer_type='v_conv'# )# model = VNet# TVNet = Encoder(#     dim=2,#     in_channels=3,#     features=(32, 64, 128, 256),#     strides=(1, 2, 2, 2),#     dilations=(1, 1, 1, 1),#     num_blocks=(2, 2, 2, 2),#     out_indices=(0, 1, 2, 3),#     layer_type='v_conv'# )# model = TVNet# ResNet18 = Encoder(#     dim=2,#     in_channels=3,#     features=(64, 128, 256, 512),#     strides=(1, 2, 2, 2),#     dilations=(1, 1, 1, 1),#     num_blocks=(2, 2, 2, 2),#     out_indices=(0, 1, 2, 3),#     first_conv=(64, 7, 2),#     layer_type='res',#     downsample=True# )# model = ResNet18# ResNet34 = Encoder(#     dim=2,#     in_channels=3,#     features=(64, 128, 256, 512),#     strides=(1, 2, 2, 2),#     dilations=(1, 1, 1, 1),#     num_blocks=(3, 4, 6, 3),#     out_indices=(0, 1, 2, 3),#     first_conv=(64, 7, 2),#     layer_type='res',#     downsample=True# )# model = ResNet34# ResNet50 = Encoder(#     dim=2,#     in_channels=3,#     features=(256, 512, 1024, 2048),#     strides=(1, 2, 2, 2),#     dilations=(1, 1, 1, 1),#     num_blocks=(3, 4, 6, 3),#     out_indices=(0, 1, 2, 3),#     first_conv=(64, 7, 2),#     layer_type='b_res',#     downsample=True# )# model = ResNet50# ResNeXt50_32x4 = Encoder(#     dim=2,#     in_channels=3,#     features=(256, 512, 1024, 2048),#     strides=(1, 2, 2, 2),#     dilations=(1, 1, 1, 1),#     num_blocks=(3, 4, 6, 3),#     out_indices=(0, 1, 2, 3),#     groups=32,#     width_per_group=4,#     first_conv=(64, 7, 2),#     layer_type='b_res',#     downsample=True# )# model = ResNeXt50_32x4# SEResNet50 = Encoder(#     dim=2,#     in_channels=3,#     features=(256, 512, 1024, 2048),#     strides=(1, 2, 2, 2),#     dilations=(1, 1, 1, 1),#     num_blocks=(3, 4, 6, 3),#     out_indices=(0, 1, 2, 3),#     first_conv=(64, 7, 2),#     layer_type='se_b_res',#     downsample=True# )# model = SEResNet50# SEResNet50 = Encoder(#     dim=2,#     in_channels=3,#     features=(256, 512, 1024, 2048),#     strides=(1, 2, 2, 2),#     dilations=(1, 1, 1, 1),#     num_blocks=(3, 4, 6, 3),#     out_indices=(0, 1, 2, 3),#     first_conv=(64, 7, 2),#     layer_type='se_b_res',#     downsample=True# )# model = SEResNet50# DeeplungResNet18 = Encoder(#     dim=3,#     in_channels=1,#     features=(32, 64, 64, 64),#     strides=(1, 2, 2, 2),#     dilations=(1, 1, 1, 1),#     num_blocks=(2, 2, 3, 3),#     out_indices=(0, 1, 2, 3),#     first_conv=(24, 7, 2),#     layer_type='res',#     groups=32,#     width_per_group=4,# )# model = DeeplungResNet18ResNet18 = Encoder(dim=3,in_channels=1,features=(16, 32, 64, 128),strides=(1, 2, 2, 2),dilations=(1, 1, 1, 1),num_blocks=(2, 2, 2, 2),out_indices=(0, 1, 2, 3),first_conv=(64, 7, 2),layer_type='res',downsample=True)model = ResNet18print(model)model.print_model_params()data = torch.ones((1, 1, 96, 96, 96))outs = model(data)for o in outs:print(o.shape)print(torch.sum(o))from medtk.runner.checkpoint import load_checkpoint# load_checkpoint(model, 'https://download.pytorch.org/models/resnet18-5c106cde.pth')# load_checkpoint(model, 'https://download.pytorch.org/models/resnet34-333f7ec4.pth')# load_checkpoint(model, 'https://download.pytorch.org/models/resnet50-19c8e357.pth')# load_checkpoint(model, 'https://download.pytorch.org/models/resnext50_32x4d-7cdf4587.pth')# load_checkpoint(model, 'http://data.lip6.fr/cadene/pretrainedmodels/se_resnet50-ce0d4300.pth')

Byobu 使用技巧相关推荐

keyshot怎么批量渲染_提高Keyshot逼真渲染的小技巧
Keyshot是一个特别神奇的应用软件,但是,就像Photoshop一样,如果你不知道怎么使用它,那么再优秀的工具在你手中也什么都是了.这里我就告诉你一些制作优秀效果图的技巧以及如何使用这个神奇软件. ...
Linux shell 学习笔记（6）— vim 编辑器使用方法及技巧
1. 检查 vim 软件包 1.1 CentOS 发行版 $ alias vi alias vi='vim' $ $ which vim /usr/bin/vim $ $ ls -l /usr/bin ...
Python 笔试面试及常用技巧（1）
1. 交换两个数字 In [66]: x, y = 1, 2In [67]: x Out[67]: 1In [68]: y Out[68]: 2 赋值的右侧形成了一个新的元组,左侧立即解析(unpac ...
Redis 使用技巧
Redis 现在非常受欢迎,似乎已经成为内存数据存储行业的标准.本人结合平时使用Redis经验,也同时查找了一些网上别人的总结经验,总结以下几条Redis使用技巧. 1. 停止使用 KEYS 众所周知 ...
受用一生的高效 PyCharm 使用技巧（六）
http://www.sohu.com/a/329854019_654419 大家好,今天我又来给大家更新 PyCharm 的使用技巧. 从第一篇开始,一直到本篇,一共更新了6篇文章,每篇 5 个小技 ...
受用一生的高效 PyCharm 使用技巧（四）
https://blog.csdn.net/pdcfighting/article/details/93269028 大家好,距离最近一篇 PyCharm 使用技巧的文章已经过去一月有余,最近虽然也比 ...
受用一生的高效 PyCharm 使用技巧（二）pycharm 指定参数运行文件
https://mp.weixin.qq.com/s/Ii0-qHUXayTPb-K-17hmQQ 在介绍技巧之前,有些话想声明一下,这个系列的一些小技巧,对于一些重试用户来说可能是小 case,如果 ...
受用一生的高效 PyCharm 使用技巧（一）
声明:本文章转自返回主页Python编程时光 PyCharm 是大多数 Python 开发者的首选 IDE,每天我们都在上面敲着熟悉的代码,写出一个又一个奇妙的功能. https://www.cnb ...
PyCharm 使用技巧
PyCharm 使用技巧 2018.12.15 00:26:36 字数 1034 阅读 290 JetBrains家的IDE很多技巧是通用的,说一些自己日常用得多但不一定仅限于PyCharm的技巧: ...

Byobu 使用技巧