--

\\n\\n

\\n
• Home
\\n
• HTML
\\n
• JavaScript
\\n
• CSS
\\n
• Vue
\\n
• React
\\n
• Python3
\\n
• Java
\\n
• C
\\n
• C++
\\n
• C#
\\n
• AI
\\n
• Go
\\n
• SQL
\\n
• Linux
\\n
• VS Code
\\n
• Bootstrap
\\n
• Git
\\n
• Local Bookmarks
\\n

\\n\\n

\\n
• Home
\\n
• HTML
\\n
• CSS
\\n
• JS
\\n
• Local Bookmarks
\\n
• Search
\\n

\\n\\n

\\n
• Vue3 Tutorial
\\n
• Vue2 Tutorial
\\n

\\n\\n

\\n
• Bootstrap3
\\n
• Bootstrap4
\\n
• Bootstrap5
\\n

\\n\\n

\\n
• Machine Learning
\\n
• PyTorch
\\n
• TensorFlow
\\n
• Sklearn
\\n
• NLP
\\n
• AI Agent
\\n
• Ollama
\\n
• Coding Plan
\\n

\\n\\n

PyTorch Tutorial

\\n\\n

\\n
• PyTorch Tutorial
\\n
• PyTorch Introduction
\\n
• PyTorch Install
\\n
• PyTorch Convolutional Neural Network
\\n
• PyTorch Recurrent Neural Network
\\n
• Pytorch torch
\\n
• PyTorch torch.nn
\\n
• Transformer Models
\\n
• PyTorch LSTM / GRU
\\n
• PyTorch Word Embeddings
\\n
• PyTorch Generative Adversarial Networks
\\n
• PyTorch Autoencoders
\\n
• PyTorch torchtext
\\n
• PyTorch Mixed Precision Training
\\n
• PyTorch TorchScript/ONNX Export
\\n
• PyTorch torch.nn Reference Manual
  \\n\\n

  PyTorch torch.triu_indices functions

  \\n\\n

  Pytorch torch Reference Manual

  \\n\\n

  torch.triu_indices

  is a function in PyTorch used to generate indices for the upper triangular matrix. It returns the row and column indices of the upper triangular part (including the diagonal).

  \\n\\n

  Function Definition

  torch.triu_indices(row, column, offset=0, dtype=torch.long, device='cpu')\\n

  \\n\\n

  Parameter Description:

  \\n\\n
  \\n
  • row: Number of Rows
  \\n
  • column: Number of Columns
  \\n
  • offset: Diagonal Offset
  \\n
  • dtype: Returned Data Type
  \\n
  • device: Device
  \\n
  \\n\\n

  ---

  \\n\\n

  Usage example

  \\n\\n

  Examples

  \\n\\n

  import torch\\n\\n# Generate Upper Triangular Indices for 3x3 Matrix\\n\\n row, col = torch.triu_indices(3,3)\\n\\nprint("row:", row)\\n\\nprint("col:", col)\\n

  The output result is:

  \\n\\n

  row: tensor([0, 0, 0, 1, 1, 2]) col: tensor([0, 1, 2, 1, 2, 2])

  \\n\\n

  Example

  import torch\\n\\n# Generate Indices and Use for Indexing Operations\\n\\n row, col = torch.triu_indices(3,3, offset=1)\\n\\n# Create 3x3 Matrix\\n\\n a = torch.ones(3,3)\\n\\n# Set Upper Triangular Part Using Indices\\n\\n a[row, col]=0\\n\\nprint(a)\\n

  The output is:

  \\n\\n

  tensor([[1., 1., 1.], [0., 1., 1.], [0., 0., 1.]])

  \\n\\n

  Example

  import torch\\n\\n# Non-square Matrix Case\\n\\n row, col = torch.triu_indices(3,4)\\n\\nprint("row:", row)\\n\\nprint("col:", col)\\n

  The output is:

  \\n\\n

  row: tensor([0, 0, 0, 1, 1, 2]) col: tensor([0, 1, 2, 1, 2, 2])

  \\n\\n

  ---

  \\n\\n

  PyTorch torch Reference Manual

  \\n\\n

  AI is thinking...

  \\n\\n

  PyTorch Data Conversion

  \\n\\n

  PyTorch torch.nn Reference Manual

  \\n\\n

  ByteArk Coding Plan supports mainstream large models such as Doubao, GLM, DeepSeek, Kimi, and MiniMax, officially provided for stable and reliable service. Configuration Guide ¥9.9/month Activate Now