Unique name : Unique name of tabs block tag without comma. Will be used in #id's as prefix for each tab with their index numbers. If there are whitespaces in name, for generate #id all whitespaces will replaced by dashes. Only for current url of post/page must be unique! [index] : Index number of active tab. If not specified, first tab (1) will be selected. If index is -1, no tab will be selected. It's will be something like spoiler. Optional parameter. [Tab caption] : Caption of current tab. If not caption specified, unique name with tab index suffix will be used as caption of tab. If not caption specified, but specified icon, caption will empty. Optional parameter. [@icon] : FontAwesome icon name (full-name, look like 'fas fa-font') Can be specified with or without space; e.g. 'Tab caption @icon' similar to 'Tab caption@icon'. Optional parameter.
今天研究生英语结课了，明天把第五组的推文发出来，我助教的工作就完全结束了，竟然有些舍不得。虽然选课时因为六级成绩过期导致无法免修英语，心里有些不甘，但对英语本身我是不排斥的，我喜欢通过英语这个媒介加强与老师和同学们的交流。一直以来口语都是我的心病，我很想得到别人的认可，却又始终不敢开口讲英语，于是My spoken English is poor这句话就常常挂在嘴边，时间久了真就觉得自己口语不行。
但是我想错了，老师拥有一套完整的方法论。从引入how to speak English in public这一主题，到passion、pronunciation、轻重缓急之区分度、sentence、structure of a paragraph、acting，循序渐进，最终带领我们呈现一个完整的英语演讲，我总能学到新的东西。每周一晚上大家一起往群里发随堂作业时，我偶尔会心血来潮把每个人的都点开看一遍，总有比我讲的好的同学，每次都会有新的体验和收获。
BIRDS（Bi-channel Image Registration and Deep-learning Segmentation for efficient, versatile mapping of mouse brain）是ImageJ的一个插件，由费鹏老师实验室开发，主要用于小鼠脑片配准和细胞计数。BIRDS改进了脑片与图谱的配准算法，并借助Imaris实现了全脑的三维可视化。
// 查看stack的某一张 // getProcessor:Returns an ImageProcessor for the specified slice, where 1<=n<=nslices. ImageProcessorimageProcessor= annotationImageStack.getProcessor(120); ImagePlusmyImg=newImagePlus("ip2", imageProcessor); myImg.show();
out = img.resize((width, height)) type = img.format out.save('D:/Desktop/elastix-5.1.0-win64/out_anno/annotationImage0270-1.tif',type) # out.save('D:/Desktop/elastix-5.1.0-win64/out_temp/tempImage0270-1.tif',type)
# 翻墙后设置代理 set http_proxy=http://127.0.0.1(代理的IP地址):1080(代理的端口号) set https_proxy=http://127.0.0.1(代理的IP地址):1080(代理的端口号) # 不翻墙的话需要取消代理，否则打开一些页面会很慢 set http_proxy= set https_proxy=
R_ABORTED: given when breaking fixation. (def: -0.1 float)
+
R_CORRECT：given when correct. (def: +1. float)
+
R_FAIL：given when incorrect. (def: 0. float)
+
+
timing：Description and duration of periods forming a trial. stim_scale：Controls the difficulty of the experiment. (def: 1., float) cohs：list of float, coherence levels controlling the difficulty of the task sigma：float, input noise level dim_ring：int, dimension of ring input and output
# import packages import time # for measuring time import numpy as np # for numerical computation from matplotlib import cm # for plotting figures import matplotlib as mpl import matplotlib.pyplot as plt from matplotlib.colors import ListedColormap, LinearSegmentedColormap from matplotlib.lines import Line2D from IPython import display # for controlling colab
import gym # for neurogym import neurogym as ngym # for training environments import torch # for training neural nets import torch.nn.functional as F
from sklearn.decomposition import PCA # for principal component analysis from scipy.special import softmax mpl.rcParams.update(mpl.rcParamsDefault)
# make supervised dataset dataset = ngym.Dataset(env, batch_size=hpar['batch_size'], seq_len=hpar['seq_len'])
# generate one batch of data when called inputs, target = dataset() print('Input to network has shape(SeqLen,Batch,Dim)=', inputs.shape) print('Target to network has shape(SeqLen,Batch)=', target.shape) # 输出 # Input to network has shape(SeqLen,Batch,Dim)= (60, 16, 3) # Target to network has shape(SeqLen,Batch)= (60, 16)
# Task structure i_trial = np.random.choice(hpar['batch_size']) # 随机从一个batch内选择一个trial end = sum([v for _,v in env.timing.items()])/1000# 单位换为s times = np.arange(end, step=env.dt/1000) rules = end - env.timing['decision']/1000, end
f, ax = plt.subplots(1, 2, figsize=(16,5)) ax[0].axvspan(rules[0], rules[1], facecolor='grey', alpha=0.2) ax[0].plot(times, inputs[:,i_trial,1], 'blue', label='Evidence for right motion') ax[0].plot(times, inputs[:,i_trial,2], 'red', label='Evidence for left motion') ax[0].plot(times, inputs[:,i_trial,0], 'green', label='Fixation rule') ax[0].set_ylabel('input') ax[0].set_ylim([-0.1,1.1]) ax[0].legend(loc='upper left') ax[0].set_xlabel('time(s)') ax[0].set_title(f'Input to RNN(trial number={i_trial})')
display.clear_output(wait=True) display.display(plt.gcf()) if i != hpar['n_iteration'] -1: cb.remove() if i == hpar['n_iteration'] -1: display.clear_output(wait=True) print(f'Final training loss={np.round(np.mean(losses[(-100):]),2)}') print(f'Final training accuracy(%)={np.round(np.mean(accs[(-100):]),2)}') return net net = RNN(input_size, output_size, **hpar) net = train_model(net, dataset, **hpar)
Part II着重介绍单个神经元动力学的简化模型。它涵盖了带有适应性和不带适应性的非线性积分-放电模型integrate-and-fire model，尤其是二次和指数的积分-放电模型，以及Izhikevich模型和自适应指数积分-发放模型。还讨论了神经动力学中噪声的问题，并呈现了两种经典的噪声描述。首先，通过随机尖峰到达引起的随机性：这个方法导致电压微分方程中的噪声项，可以用朗之万方程表示。其次，神经元内在的随机性导致在亚阈区域时跨越触发阈值的“逃逸”：这个方法引出了广义线性模型的框架，该模型会在神经编码和解码的应用中系统地介绍和讨论。强调第二部分中神经元模型和生物数据之间的关系，并介绍了系统的参数优化算法。
+
Part III将第二部分得出的简化模型用于构建网络。以群体活动方程，也称为群体发放率，描述神经网络动力学的集体特性。确定了可以使用标准发放速率模型描述群体活动的条件。
+
Part IV将动力学与认知联系起来。使用群体活动方程分析计算和认知神经科学中著名的范式，比如在决策或记忆检索过程中的神经活动。在第四部分，我们也概述了与突触可塑性相关的学习理论。本书以极有吸引力的神经动力学原理帮助帕金森病人的应用结尾。
在github下载示例插件并修改 GitHub - imagej/example-imagej2-command: Simple Maven project for an ImageJ2 command
+
+
+
修改pom.xml：重点修改groupId、artifactId、version、name、的内容
+
Eclipse-Import-Maven-Existing Maven Projects
+
重命名：项目名、包名、java文件中类名和菜单路径
+
+
+
Run as Java application 确保程序可以正常运行
+
Run as Maven build 控制台输出BUILD SUCCESS 则表示程序成功打包为jar包（保存在target目录下）
+
将jar包粘贴至Fiji的plugin目录，运行软件测试插件效果
+
+
问题记录
+
如果pom.xml中报错，则尝试修改，并等待右下角更新完成
+
打包好的插件在ImageJ中无法正常运行，报错java.lang.ClassCastException: com.chen.brain.pluginMain cannot be cast to org.scijava.plugin.SciJavaPlugin 报错原因：pluginMain类没有实现org.scijava.command.Command接口，而这是必须的，才能将其识别为有效的SciJava插件。@Plugin注解用于指定该类应注册为一个命令插件，但由于它没有实现所需的接口，因此会导致ClassCastException异常。 解决方案：
Error loading webview: Error: Could not register service worker: InvalidStateError: Failed to register a ServiceWorker: The document is in an invalid state..
+
The editor could not be opened due to an unexpected error: Could not initialize webview: Error: Could not register service worker: InvalidStateError: Failed to register a ServiceWorker: The document is in an invalid state..
void setup() { // put your setup code here, to run once: pinMode(buttonPin, INPUT_PULLUP); pinMode(ledPin, OUTPUT); }
void loop() { // put your main code here, to run repeatedly: if (digitalRead(buttonPin) == LOW){ digitalWrite(ledPin, HIGH); } else{ digitalWrite(ledPin, LOW); } }
参考文献1 We first examined whether a neuron was excited or inhibited by light stimulation by asking whether its firing rate during the first 500 ms after light onset was higher or lower, respectively, than its firing rate in the 500 ms immediately before light onset. This was examined separately for each light intensity (Wilcoxon signed-rank test, p < 0.05, corrected for multiple comparisons). In order to quantify the strength of inhibition we calculated the normalized firing rate by dividing the mean firing in the first 500 ms after laser onset by the mean baseline firing in the 500 ms before laser onset.
+
Babl S S, Rummell B P, Sigurdsson T. The spatial extent of optogenetic silencing in transgenic mice expressing channelrhodopsin in inhibitory interneurons[J]. Cell reports, 2019, 29(5): 1381-1395. e4.
+
+
参考文献2 Food-cue, lever-press, and dish-entry responses were calculated as Z-scores normalized to 20 pre-cue bins of 300 ms. Neurons showing a Z-score >2.58 (P < 0.01) during the first two bins following the onset of the food cue, lever press, or dish entry were classified as excitatory responses, whereas neurons showing a Z-score < −1.96 (P < 0.05) during the same first two bins were classified as inhibitory responses.
+
Engelke D S, Zhang X O, O’Malley J J, et al. A hypothalamic-thalamostriatal circuit that controls approach-avoidance conflict in rats[J]. Nature communications, 2021, 12(1): 2517.
+
+
神经元类型
The spike width, defined as the time from the valley to the return to baseline after the peak, and the ratio between peak and valley amplitudes.
#include <Adafruit_GFX.h> // Core graphics library by Adafruit #include <Arduino_ST7789.h> // Hardware-specific library for ST7789 (with or without CS pin) #include <SPI.h>
#define TFT_DC 8 #define TFT_RST 7 //9 #define TFT_CS 9 //10 // only for displays with CS pin #define TFT_MOSI 11 // for hardware SPI data pin (all of available pins) #define TFT_SCLK 13 // for hardware SPI sclk pin (all of available pins)
安装方式：在Arduino的项目-加载库-库管理器中搜索GFX Library For Arduino，选择最新版本进行安装。Github下载地址 运行画直线的程序，用时约5.7s，速度超级快！在测试Arduino_GFX的时候还发生了一个小插曲，我连线明明是正确的，但屏幕没有任何反应，排了很久的bug，最后发现是我的arduino开发板有问题，果然太便宜了容易出问题，以后还是用正版的arduino开发板吧！
#include <Adafruit_GFX.h> // Core graphics library by Adafruit #include <Arduino_ST7789.h> // Hardware-specific library for ST7789 (with or without CS pin) #include <SPI.h>
#define TFT_DC 8 #define TFT_RST 7 //9 #define TFT_CS 9 //10 // only for displays with CS pin #define TFT_MOSI 11 // for hardware SPI data pin (all of available pins) #define TFT_SCLK 13 // for hardware SPI sclk pin (all of available pins)
安装方式：在Arduino的项目-加载库-库管理器中搜索GFX Library For Arduino，选择最新版本进行安装。Github下载地址 运行画直线的程序，用时约5.7s，速度超级快！在测试Arduino_GFX的时候还发生了一个小插曲，我连线明明是正确的，但屏幕没有任何反应，排了很久的bug，最后发现是我的arduino开发板有问题，果然太便宜了容易出问题，以后还是用正版的arduino开发板吧！
参考文献1 We first examined whether a neuron was excited or inhibited by light stimulation by asking whether its firing rate during the first 500 ms after light onset was higher or lower, respectively, than its firing rate in the 500 ms immediately before light onset. This was examined separately for each light intensity (Wilcoxon signed-rank test, p < 0.05, corrected for multiple comparisons). In order to quantify the strength of inhibition we calculated the normalized firing rate by dividing the mean firing in the first 500 ms after laser onset by the mean baseline firing in the 500 ms before laser onset.
Babl S S, Rummell B P, Sigurdsson T. The spatial extent of optogenetic silencing in transgenic mice expressing channelrhodopsin in inhibitory interneurons[J]. Cell reports, 2019, 29(5): 1381-1395. e4.
参考文献2 Food-cue, lever-press, and dish-entry responses were calculated as Z-scores normalized to 20 pre-cue bins of 300 ms. Neurons showing a Z-score >2.58 (P < 0.01) during the first two bins following the onset of the food cue, lever press, or dish entry were classified as excitatory responses, whereas neurons showing a Z-score < −1.96 (P < 0.05) during the same first two bins were classified as inhibitory responses.
Engelke D S, Zhang X O, O’Malley J J, et al. A hypothalamic-thalamostriatal circuit that controls approach-avoidance conflict in rats[J]. Nature communications, 2021, 12(1): 2517.
神经元类型
The spike width, defined as the time from the valley to the return to baseline after the peak, and the ratio between peak and valley amplitudes.
void setup() { // put your setup code here, to run once: pinMode(buttonPin, INPUT_PULLUP); pinMode(ledPin, OUTPUT); }
void loop() { // put your main code here, to run repeatedly: if (digitalRead(buttonPin) == LOW){ digitalWrite(ledPin, HIGH); } else{ digitalWrite(ledPin, LOW); } }
Error loading webview: Error: Could not register service worker: InvalidStateError: Failed to register a ServiceWorker: The document is in an invalid state..
The editor could not be opened due to an unexpected error: Could not initialize webview: Error: Could not register service worker: InvalidStateError: Failed to register a ServiceWorker: The document is in an invalid state..
在github下载示例插件并修改 GitHub - imagej/example-imagej2-command: Simple Maven project for an ImageJ2 command
修改pom.xml：重点修改groupId、artifactId、version、name、的内容
Eclipse-Import-Maven-Existing Maven Projects
重命名：项目名、包名、java文件中类名和菜单路径
Run as Java application 确保程序可以正常运行
Run as Maven build 控制台输出BUILD SUCCESS 则表示程序成功打包为jar包（保存在target目录下）
将jar包粘贴至Fiji的plugin目录，运行软件测试插件效果
问题记录
如果pom.xml中报错，则尝试修改，并等待右下角更新完成
打包好的插件在ImageJ中无法正常运行，报错java.lang.ClassCastException: com.chen.brain.pluginMain cannot be cast to org.scijava.plugin.SciJavaPlugin 报错原因：pluginMain类没有实现org.scijava.command.Command接口，而这是必须的，才能将其识别为有效的SciJava插件。@Plugin注解用于指定该类应注册为一个命令插件，但由于它没有实现所需的接口，因此会导致ClassCastException异常。 解决方案：
作者：Bartolo R, Averbeck B B.标题：Prefrontal cortex predicts state switches during reversal learning期刊：Neuron, 2020
+
+具体内容研究目的/科学问题研究背景研究结果研究结论想法记录可以跟进的参考文献可以借鉴的内容其它
#include <Adafruit_GFX.h> // Core graphics library by Adafruit #include <Arduino_ST7789.h> // Hardware-specific library for ST7789 (with or without CS pin) #include <SPI.h>
#define TFT_DC 8 #define TFT_RST 7 //9 #define TFT_CS 9 //10 // only for displays with CS pin #define TFT_MOSI 11 // for hardware SPI data pin (all of available pins) #define TFT_SCLK 13 // for hardware SPI sclk pin (all of available pins)
安装方式：在Arduino的项目-加载库-库管理器中搜索GFX Library For Arduino，选择最新版本进行安装。Github下载地址 运行画直线的程序，用时约5.7s，速度超级快！在测试Arduino_GFX的时候还发生了一个小插曲，我连线明明是正确的，但屏幕没有任何反应，排了很久的bug，最后发现是我的arduino开发板有问题，果然太便宜了容易出问题，以后还是用正版的arduino开发板吧！
参考文献1 We first examined whether a neuron was excited or inhibited by light stimulation by asking whether its firing rate during the first 500 ms after light onset was higher or lower, respectively, than its firing rate in the 500 ms immediately before light onset. This was examined separately for each light intensity (Wilcoxon signed-rank test, p < 0.05, corrected for multiple comparisons). In order to quantify the strength of inhibition we calculated the normalized firing rate by dividing the mean firing in the first 500 ms after laser onset by the mean baseline firing in the 500 ms before laser onset.
Babl S S, Rummell B P, Sigurdsson T. The spatial extent of optogenetic silencing in transgenic mice expressing channelrhodopsin in inhibitory interneurons[J]. Cell reports, 2019, 29(5): 1381-1395. e4.
参考文献2 Food-cue, lever-press, and dish-entry responses were calculated as Z-scores normalized to 20 pre-cue bins of 300 ms. Neurons showing a Z-score >2.58 (P < 0.01) during the first two bins following the onset of the food cue, lever press, or dish entry were classified as excitatory responses, whereas neurons showing a Z-score < −1.96 (P < 0.05) during the same first two bins were classified as inhibitory responses.
Engelke D S, Zhang X O, O’Malley J J, et al. A hypothalamic-thalamostriatal circuit that controls approach-avoidance conflict in rats[J]. Nature communications, 2021, 12(1): 2517.
神经元类型
The spike width, defined as the time from the valley to the return to baseline after the peak, and the ratio between peak and valley amplitudes.
void setup() { // put your setup code here, to run once: pinMode(buttonPin, INPUT_PULLUP); pinMode(ledPin, OUTPUT); }
void loop() { // put your main code here, to run repeatedly: if (digitalRead(buttonPin) == LOW){ digitalWrite(ledPin, HIGH); } else{ digitalWrite(ledPin, LOW); } }
Error loading webview: Error: Could not register service worker: InvalidStateError: Failed to register a ServiceWorker: The document is in an invalid state..
The editor could not be opened due to an unexpected error: Could not initialize webview: Error: Could not register service worker: InvalidStateError: Failed to register a ServiceWorker: The document is in an invalid state..
在github下载示例插件并修改 GitHub - imagej/example-imagej2-command: Simple Maven project for an ImageJ2 command
修改pom.xml：重点修改groupId、artifactId、version、name、的内容
Eclipse-Import-Maven-Existing Maven Projects
重命名：项目名、包名、java文件中类名和菜单路径
Run as Java application 确保程序可以正常运行
Run as Maven build 控制台输出BUILD SUCCESS 则表示程序成功打包为jar包（保存在target目录下）
将jar包粘贴至Fiji的plugin目录，运行软件测试插件效果
问题记录
如果pom.xml中报错，则尝试修改，并等待右下角更新完成
打包好的插件在ImageJ中无法正常运行，报错java.lang.ClassCastException: com.chen.brain.pluginMain cannot be cast to org.scijava.plugin.SciJavaPlugin 报错原因：pluginMain类没有实现org.scijava.command.Command接口，而这是必须的，才能将其识别为有效的SciJava插件。@Plugin注解用于指定该类应注册为一个命令插件，但由于它没有实现所需的接口，因此会导致ClassCastException异常。 解决方案：
Part II着重介绍单个神经元动力学的简化模型。它涵盖了带有适应性和不带适应性的非线性积分-放电模型integrate-and-fire model，尤其是二次和指数的积分-放电模型，以及Izhikevich模型和自适应指数积分-发放模型。还讨论了神经动力学中噪声的问题，并呈现了两种经典的噪声描述。首先，通过随机尖峰到达引起的随机性：这个方法导致电压微分方程中的噪声项，可以用朗之万方程表示。其次，神经元内在的随机性导致在亚阈区域时跨越触发阈值的“逃逸”：这个方法引出了广义线性模型的框架，该模型会在神经编码和解码的应用中系统地介绍和讨论。强调第二部分中神经元模型和生物数据之间的关系，并介绍了系统的参数优化算法。
Part III将第二部分得出的简化模型用于构建网络。以群体活动方程，也称为群体发放率，描述神经网络动力学的集体特性。确定了可以使用标准发放速率模型描述群体活动的条件。
Part IV将动力学与认知联系起来。使用群体活动方程分析计算和认知神经科学中著名的范式，比如在决策或记忆检索过程中的神经活动。在第四部分，我们也概述了与突触可塑性相关的学习理论。本书以极有吸引力的神经动力学原理帮助帕金森病人的应用结尾。
R_ABORTED: given when breaking fixation. (def: -0.1 float)
R_CORRECT：given when correct. (def: +1. float)
R_FAIL：given when incorrect. (def: 0. float)
timing：Description and duration of periods forming a trial. stim_scale：Controls the difficulty of the experiment. (def: 1., float) cohs：list of float, coherence levels controlling the difficulty of the task sigma：float, input noise level dim_ring：int, dimension of ring input and output
# import packages import time # for measuring time import numpy as np # for numerical computation from matplotlib import cm # for plotting figures import matplotlib as mpl import matplotlib.pyplot as plt from matplotlib.colors import ListedColormap, LinearSegmentedColormap from matplotlib.lines import Line2D from IPython import display # for controlling colab
import gym # for neurogym import neurogym as ngym # for training environments import torch # for training neural nets import torch.nn.functional as F
from sklearn.decomposition import PCA # for principal component analysis from scipy.special import softmax mpl.rcParams.update(mpl.rcParamsDefault)
# make supervised dataset dataset = ngym.Dataset(env, batch_size=hpar['batch_size'], seq_len=hpar['seq_len'])
# generate one batch of data when called inputs, target = dataset() print('Input to network has shape(SeqLen,Batch,Dim)=', inputs.shape) print('Target to network has shape(SeqLen,Batch)=', target.shape) # 输出 # Input to network has shape(SeqLen,Batch,Dim)= (60, 16, 3) # Target to network has shape(SeqLen,Batch)= (60, 16)
# Task structure i_trial = np.random.choice(hpar['batch_size']) # 随机从一个batch内选择一个trial end = sum([v for _,v in env.timing.items()])/1000# 单位换为s times = np.arange(end, step=env.dt/1000) rules = end - env.timing['decision']/1000, end
f, ax = plt.subplots(1, 2, figsize=(16,5)) ax[0].axvspan(rules[0], rules[1], facecolor='grey', alpha=0.2) ax[0].plot(times, inputs[:,i_trial,1], 'blue', label='Evidence for right motion') ax[0].plot(times, inputs[:,i_trial,2], 'red', label='Evidence for left motion') ax[0].plot(times, inputs[:,i_trial,0], 'green', label='Fixation rule') ax[0].set_ylabel('input') ax[0].set_ylim([-0.1,1.1]) ax[0].legend(loc='upper left') ax[0].set_xlabel('time(s)') ax[0].set_title(f'Input to RNN(trial number={i_trial})')
display.clear_output(wait=True) display.display(plt.gcf()) if i != hpar['n_iteration'] -1: cb.remove() if i == hpar['n_iteration'] -1: display.clear_output(wait=True) print(f'Final training loss={np.round(np.mean(losses[(-100):]),2)}') print(f'Final training accuracy(%)={np.round(np.mean(accs[(-100):]),2)}') return net net = RNN(input_size, output_size, **hpar) net = train_model(net, dataset, **hpar)
# 翻墙后设置代理 set http_proxy=http://127.0.0.1(代理的IP地址):1080(代理的端口号) set https_proxy=http://127.0.0.1(代理的IP地址):1080(代理的端口号) # 不翻墙的话需要取消代理，否则打开一些页面会很慢 set http_proxy= set https_proxy=
out = img.resize((width, height)) type = img.format out.save('D:/Desktop/elastix-5.1.0-win64/out_anno/annotationImage0270-1.tif',type) # out.save('D:/Desktop/elastix-5.1.0-win64/out_temp/tempImage0270-1.tif',type)
// 查看stack的某一张 // getProcessor:Returns an ImageProcessor for the specified slice, where 1<=n<=nslices. ImageProcessorimageProcessor= annotationImageStack.getProcessor(120); ImagePlusmyImg=newImagePlus("ip2", imageProcessor); myImg.show();
BIRDS（Bi-channel Image Registration and Deep-learning Segmentation for efficient, versatile mapping of mouse brain）是ImageJ的一个插件，由费鹏老师实验室开发，主要用于小鼠脑片配准和细胞计数。BIRDS改进了脑片与图谱的配准算法，并借助Imaris实现了全脑的三维可视化。
但是我想错了，老师拥有一套完整的方法论。从引入how to speak English in public这一主题，到passion、pronunciation、轻重缓急之区分度、sentence、structure of a paragraph、acting，循序渐进，最终带领我们呈现一个完整的英语演讲，我总能学到新的东西。每周一晚上大家一起往群里发随堂作业时，我偶尔会心血来潮把每个人的都点开看一遍，总有比我讲的好的同学，每次都会有新的体验和收获。
Unique name : Unique name of tabs block tag without comma. Will be used in #id's as prefix for each tab with their index numbers. If there are whitespaces in name, for generate #id all whitespaces will replaced by dashes. Only for current url of post/page must be unique! [index] : Index number of active tab. If not specified, first tab (1) will be selected. If index is -1, no tab will be selected. It's will be something like spoiler. Optional parameter. [Tab caption] : Caption of current tab. If not caption specified, unique name with tab index suffix will be used as caption of tab. If not caption specified, but specified icon, caption will empty. Optional parameter. [@icon] : FontAwesome icon name (full-name, look like 'fas fa-font') Can be specified with or without space; e.g. 'Tab caption @icon' similar to 'Tab caption@icon'. Optional parameter.
+
+
+ 
+ 
+
+
+
+
+
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
+
+
+
+
+
+
+
+
+
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+