- Deep Learning with PyTorch
- Vishnu Subramanian
- 252字
- 2021-06-24 19:16:23
Slicing tensors
A common thing to do with a tensor is to slice a portion of it. A simple example could be choosing the first five elements of a one-dimensional tensor; let's call the tensor sales. We use a simple notation, sales[:slice_index] where slice_index represents the index where you want to slice the tensor:
sales = torch.FloatTensor([1000.0,323.2,333.4,444.5,1000.0,323.2,333.4,444.5])
sales[:5]
1000.0000
323.2000
333.4000
444.5000
1000.0000
[torch.FloatTensor of size 5]
sales[:-5]
1000.0000
323.2000
333.4000
[torch.FloatTensor of size 3]
Let's do more interesting things with our panda image, such as see what the panda image looks like when only one channel is chosen and see how to select the face of the panda.
Here, we select only one channel from the panda image:
plt.imshow(panda_tensor[:,:,0].numpy())
#0 represents the first channel of RGB
The output is as follows:

Now, lets crop the image. Say we want to build a face detector for pandas and we need just the face of a panda for that. We crop the tensor image such that it contains only the panda's face:
plt.imshow(panda_tensor[25:175,60:130,0].numpy())
The output is as follows:

Another common example would be where you need to pick a specific element of a tensor:
#torch.eye(shape) produces an diagonal matrix with 1 as it diagonal #elements.
sales = torch.eye(3,3)
sales[0,1]
Output- 0.00.0
We will revisit image data in Chapter 5, Deep Learning for Computer Vision, when we discuss using CNNs to build image classifiers.
- 計算機組裝與系統配置
- 電腦常見故障現場處理
- 電腦組裝、維護、維修全能一本通(全彩版)
- 嵌入式系統設計教程
- Artificial Intelligence Business:How you can profit from AI
- Practical Machine Learning with R
- 電腦高級維修及故障排除實戰
- 固態存儲:原理、架構與數據安全
- Managing Data and Media in Microsoft Silverlight 4:A mashup of chapters from Packt's bestselling Silverlight books
- Blender for Video Production Quick Start Guide
- 嵌入式系統設計大學教程(第2版)
- Advanced Machine Learning with R
- 筆記本電腦現場維修實錄
- 施耐德M241/251可編程序控制器應用技術
- Service Mesh微服務架構設計