Kaggle Introduction

•

Kaggle.com 人工智慧競賽的雲端平台, 提供免費使用 CPU / GPU / TPU

Kaggle.com

人工智慧競賽的雲端平台, 提供免費的CPU / GPU / TPU

  • CPU: Intel Xeon 2.0GHz, 1 core - 2 threads
  • GPU: Nvidia Tesla P100-PCIE-16GB
  • TPU: Google TPU v3-8 (8 cores)

Sign-In

  • Register> Register with Google
  • Sign-In> Sign In with Google
  • kaggle.com/yourname

Code (程式碼)

+ New Notebook

  • rename notebook
  • delete cell
  • Settings (Internet on)
  • key-in command !lscpu
  • session operations
  • start session
  • Run> Run current Cell
  • Restart & Clear Cell Outputs
  • Edit & Run
  • + Code
  • Accelerator
  • Turn on GPU
  • Run All
  • Save Version
  • Save & Run All (Commit) Save
  • Go to Viewer
  • Share>
  • Share> Public Save

Dataset (資料集)

  • Datasets
  • + New Dataset
  • keyin name of Dataset, and set to Public
  • Drag and drop files Create
  • Go to Dataset

Notebook (筆記本)

  • Search [kaggle.com/rkuo2000/code](https://kaggle.com/rkuo2000/code]

  • Easy Net Copy & Edit
  • Run All

Import Notebook

  • File> Import Notebook
  • drag file to import

Neural Network (神經網路)

Easy Net

1 neuron to learn Linear equation: Y= 0.1*X +0.3

Hidden Net

10 neurons to learn parabola equestion: Y = X^2



This site was last updated December 22, 2022.