Wu Times

1. conda create -n <env_name> <python=3.9>

2. conda install jupyter notebook

3. pip install jupyter_contrib_nbextensions, jupyter contrib nbextension install --user, pip install jupyter_nbextensions_configurator, jupyter nbextensions_configurator enable --user

4. open jupyter notebook->Nbextension->Hinterland (enable this button)

5. conda install nb_conda_kernels

   with this package, you are able to select your own conda environment in notebook.

6. Click the method/variable in jupyter and press shift+tab to view the detailed information.

Install docker

Download docker here: https://docs.docker.com/docker-for-mac/apple-m1/

Note:

This is just a preview version, which means you may encounter various bugs.

This version is based on Rosetta2, so remember to install it via $ softwareupdate --install-rosetta.

Check your docker

1
2

$ docker -v
$ docker info // it may take a few minutes to start docker

Use docker image MIT 6.S081

1
2

$ docker pull calvinhaynes412/mit6.s081:v1 // pull docker image
$ git clone git@github.com:CalvinHaynes/MIT6.S081-2020Fall.git // pull git repository

Open the docker image->set repository to the local git repository->open docker in browser.

1. Groundwater(地下水) is the word used to describe water that saturates(渗透、浸透) the ground.

2. By far(到目前为止) the most abundant(大量的、充足的) type of groundwater is meteoric(大气的；流星的) water.

3. circulate(v. 循环；流通；流传)

4. precipitation(化学沉淀；降水；冰雹；坠落)

5. sand grains(沙粒)--grain(谷物；颗粒)

尝试了NPR的场景渲染，感觉不是特别成功，二次元渲染不仅需要技术，调参也很难，需要一些美学功底更重要的是对渲染节点的深入理解。

初次尝试了blender里钻石材质的生成，接下来会详细分析材质节点的组成并仔细调参。并且会更换钻石模型和适当的场景进行测试。

Prefab

本文记录了我自己在刷剑指offer时的代码。题目顺序根据牛客网，代码没有极致追求高效，但是尽可能使用优雅简明的实现方式，但同时也不会为了简单而使用大量STL自带的方法（若使用，尽可能给出具体实现方式）。

Prefab

Games101还是要强烈推荐给大家，把图形学一些基本的内容讲的十分透彻了。这里记录一下作业的成果，具体的代码实现就不展示了，如果有特别巧妙的会注明。封面将对当前最新进度结果进行展示。

已完结，代码库

Precision

\(Precision=\frac{actually \ true \&\& model \ thinks \ it \ is \ true}{model \ thinks \ it \ is \ true}\)

The percentage of true labels decided by the model are actually true.

Recall

\(Recall=\frac{model \ thinks \ it \ is \ true \&\&actually \ true}{actually \ true}\)

The percentage of true items are decided true by the model.

If we want higher Precision, we should not take any risk--only take those with highest score.

But if we want higher Recall, we should take all possible results. That is, we cannot get Precision and Recall improved at the same time.

So we need a balance benchmark.

F1-Measure/F1-Score

\[ F=\frac{(\alpha ^2+1)PR}{\alpha ^2P+R} \]

We get F1-Measure when \(\alpha=1\).

More on Harmonic Mean--it prevents models that are only good at either of P or R from doing good in F.

x=50, y=0~100, here is the score (ref):

Harmonic mean tries to get a result that neither P nor R is not so bad even if both of them are not really good.

Conda Environment

1. Because anaconda is not suitable for M1, we use miniforge3 instead.
2. Create new conda environment with 'python=3.9'. (miniforge3 only support python3.9 on M1).

PyQt5

1. install homebrew (remeber to add path).
2. brew install pyqt5. Homebrew supports pyqt now.
3. add path: echo 'export PATH="/opt/homebrew/opt/qt@5/bin:$PATH"' >> ~/.bashrc and echo 'export PATH="/opt/homebrew/opt/pyqt@5/bin:$PATH"' >> ~/.bashrc. This path may be not suitbale for you. You should modify your path to /qt5@/bin and /qypt@5/bin, and modify your bash profile (~/.zshrc for most Macbook users).
4. Duplicate the files in /opt/homebrew/Cellar/pyqt@5/5.15.4_1/lib/python3.9/site-packages and paste them to /Users/<username>/miniforge3/envs/<environmentname>/lib/python3.9/site-packages.
5. source ~/.bashrc (.zshrc)

labelme

1. labelme can be installed directly via pip (in your conda environemnt)
2. pip install labelme
3. when everything is done, labelme to use it.

Linear Regression

Setup

Data: \(D={(x _1,y _1),(x _2,y _2), \cdots ,(x _N,y _N) }\), \(x _i \in \pmb{R} ^p\) and \(y _i \in \pmb{R}\).

In matrix form:

​ Input: \(X=(x _1, x _2, \cdots, x _N)^T \in \pmb{R} _{N\times p}\) is a matrix. \(x _1=(x _{11},x _{12},\cdots,x _{1p})\in \pmb{R} _p\) is a vector.

​ Output: \(Y=(y _1,y _2, \cdots, y _N) \in \pmb{R} _N\) is a matrix.

​ Weight: \(W=(w _1, w _2, \cdots, w _p) ^T\) is a vector.