阿什顿粉
申达股份
sad
生地方阿什顿粉阿什顿
hh
5671234666
是打发
-
- 天凉了,记得加衣喔
- [ ] - [ ]
- [ ]
- [ ]
-
[x]
- [ ] -
- [ ]
-
[ ]
- [ ]
- [ ]
-
[ ]
1 | 1 | |
2 | ||
st=>start: 开始
e=>end: 结束
op=>operation: 我的操作
cond=>condition: 判断确认?
st->op->cond
cond(yes)->e
cond(no)->op
\(ondition\)
生
的
反攻倒算
sdfg
人生自古谁
![](Excel 2024-01-09 19.53.28.sheet)
[[../萨达发顺丰asdf.md]]
[[../萨达发顺丰asdf.md]]
是大法官
4
3
4
\(undefined\)
\(undefined\)
申达股
申达
申达
申达股
阿什顿粉
![](Drawing 2024-01-09 16.02.21.excalidraw)
-
[ ]
-
[ ]
-
[ ]
-
D
Sadfasdf
-[]
-
[]
过去 | 过去将来 | 现在 | 将来 | |
---|---|---|---|---|
一般 | did | would do | do | will do |
进行 | was/were doing | would be doing | am/is/are doing | will be doing |
完成 | had done | would have done | have done | will have done |
完成进行 | had been doing | would have been doing | have been doing | will have been doing |
sdf
萨达发sdf
[!note] 萨达发
萨达发
萨达发
[!note] 是大法师
阿什顿粉
萨达发
[!success] 萨达发
sdf
Open: Pasted image 20240111205326.png
df | s df | sdf |
---|---|---|
print(2+3)
[2table]
- 测试1|测试2
- 测试3|测试4
public class ObTest{
public static void main(String args[]) {
System.out.println("hello world!");
}
}
过去 | 过去将来 | |
---|---|---|
:--- | :---:--- | :---:--- |
一般 | did | would do |
进行 | was/were doing | would be doing |
完成 | had done | would have done |
完成进行 | had been doing | would have been doing |
Open: Pasted image 20240111205415.png | ||
Open: 38f24b2a4b47d37025091ddf2c9f6d93610x14989.jpg
人生
人的
人生
file.name
人生
人生
人生
是大
人生
人
生自古谁
人生
would have been doing
xjl xj
xjl456852
人生自古谁
人生
人
人
人生
人
xjl
x
regression
人生
2
人
完成
1 | sdf | sdf | ||
---|---|---|---|---|
2 | asd | |||
3 | ||||
>>>sum() | ||||
过去 | 过去将来 | 现在 | 将来 | |
:--- | :---:--- | :---:--- | :---:--- | :---:--- |
一般 | did | would do | do | will do |
进行 | was/were doing | would be doing | am/is/are doing | will be doing |
完成 | had done | would have done | have done | will have done |
完成进行 | had been doing | would have been doing | have been doing | will have been doing |
框架
过去 | 过去将来 | 现在 | 将来 | |
---|---|---|---|---|
一般 | did | would do | do | will do |
进行 | was/were doing | would be doing | am/is/are doing | will be doing |
完成 | had done | would have done | have done | will have done |
完成进行 | had been doing | would have been doing | have been doing | will have been doing |
- 基于最基本的线性回归模型p
- 和MSE 损失函数
- 基于EQ1 和 EQ2可以推导N个样本的 MSE Loss :
\(l_{mse}=\frac{1}{n}\sum_{j=1}^{n}l_{j}=\frac{1}{n}\sum_{j=1}^{n}(y_{j}-y_{j_{-}pred})^{2}=\frac{1} {n}\sum_{j=1}^{n}[y_{j}-(W*X_{j}+b)]^{2} --EQ(3)\)
- 基于EQ3,可以推导出Loss对W 和b的偏导表达式:
\(\frac{dl_{mse}}{dW}=\frac{2}{n}\sum_{j=1}^{n}[-X_{j}(y_{j}-(W*X_{j}+b))] -- EQ(4)\)
\(\frac{dl_{mse}}{db}=\frac{2}{n}\sum_{j=1}^{n}[-(y_{j}-(W*X_{j}+b))] --EQ(5)\)
\(\frac{1}{1}\)
- 最后是W和b的迭代公式:
\(W_{update}=W_{old}-lr*\sum_{j=1}^{n}\frac{dl_{j}}{dW} -- EQ (6)\)
\(b_{update}=b_{old}-lr*\sum_{j=1}^{n}\frac{dl_{j}}{db} -- EQ(7)\)
人生自古谁
人生自古谁无死,留取丹心照汗青
xjl456852