机器学习第二周周报

一、Transforms

1.Transforms的结构与用法

from PIL import Image
from torchvision import transforms

img = image.open(img.path)

tensor_trans = transforms.ToTensor()
tensor_img = tensor.trans(img)

from PIL import Image
from torch.utils.tensorboard import SummaryWriter
from torchvision import transforms

img = image.open(img.path)

writer = SummaryWriter("logs")

tensor_trans = transforms.ToTensor()
tensor_img = tensor.trans(img)

writer.add_image("Tensor_img",tensor_img)

writer.close()

2.常见的Transforms

（1）关注点

• 输入
• 输出
• 作用

（2）代码

from PIL import Image
from torch.utils.tensorboard import SummaryWriter
from torchvision import transforms

img = image.open(img.path)

writer = SummaryWriter("logs")

# ToTensor
tensor_trans = transforms.ToTensor()
tensor_img = tensor.trans(img)
writer.add_image("Tensor_img",tensor_img)

# Normalize
trans_more = transforms.Normalize([0.5,0.5,0.5],[0.5,0.5,0.5])
img_norm = trans_norm(img_tensor)
print(img_norm[0][0][0])
writer.add_image("Normalize",img_norm,2)

# Resize
print(img.size)
trans_resize = transforms.Resize((512,512))
# img PIL -> resize -> img_resize PIL
img_resize = trans_reszie(img)
# img_resize PIL -> totensor ->img_resize tensor
img_resize = trans_totensor(img_resize)
writer.add_image("Resize",img_resize,0)
print(img_resize)

# Compose -resize -2
trans_reszie_2 = transforms.Resize(512)
# PIL -> PIL -> tensor
trans_compose = transforms.Compose([trans_resize_2,trans_totensor])
img_resize_2 = trans_compose(img)
writer.add_image("Resize",img_resize_2,1)

#RandomCrop
trans_random = transforms.RandomCrop(500,1000)
trans_compose_2 = transforms_Compose([trans_random,trans_totensor])
for i in range(10):
	img_crop = trans_Compose_2(img)
	writer.add_image("RandomCrop",img_crop,i)

writer.close()

class Person:
	def __call__(self,name):
		print("__call__" + "hello" + name)

	def hello(self,name):
		print("hello" + name)

person = Person()
person("zhangsan")
person.hello("lisi")

二、Google Colab

1.What is Colab?

Colab, or “Colaboratory”, allows you to write and execute Python in your
browser, with

• Zero configuration required
• Free access to GPUs
• Easy sharing

2.Useful Linux Commands (in Colab)

• ls : List all files in the current directory
• ls -l : List all files in the current directory with more detail
• pwd : Output the working directory
• mkdir <dirname>: Create a directory <dirname>
• cd <dirname> : Move to directory <dirname>
• gdown : Download files from google drive
• wget : Download files from the internet
• python <python_file>: Executes a python file

三、深度学习

1.Three Steps for Deep Learning

• Step 1: Neural Network
• Step 2: goodness of function
• Step 3: pick the best function

2.Neural Network

Different connection leads to different network structures.

Network parameter 𝜃: all the weights and biases in the “neurons”

3.goodness of function

4.pick the best function

四、Backpropagation

To compute the gradients efficiently, we use backpropagation.

1.Forward pass

Compute 𝜕𝑧 / 𝜕𝑤 for all parameters

2.Backward pass

Compute 𝜕𝐶 / 𝜕𝑧 for all activation function inputs z




𝜎′ (z) is a constant because z is
already determined in the forward pass.

3.Summary

五、预测神奇宝贝Pokemon

1.Step 1: Model

y = b +xcp
(w and b are parameters (can be any value))
Linear model: y = b + Σwixi

2.Step 2: Goodness of Function

3.Step 3: Best Function

最终形成一下等高线图

4.result

使用训练集和测试集的平均误差来验证模型：

5.Generalization

优化上述模型：

A more complex model does not always lead to better performance on testing data.
This is Overfitting. => Select suitable model

6.考虑其他的因素

我们需要考虑其他的属性是否会影响最终的结果。

7.重新定义function

（1）Back to step 1: Redesign the Model

（2）Back to step 2: Regularization

• Training error: larger𝜆, considering the training error less
• We prefer smooth function, but don’t be too smooth.