This project utilizes the Jittor framework to train a Conditional Generative Adversarial Network (CGAN) model on the MNIST digital image dataset.
By providing a random vector z and additional auxiliary information 𝑦 (such as class labels), the model generates images of specific digits.
The resulting generated image corresponds to the random ID "20349262030119" (see result.png).
The overall accuracy achieved is 0.9709.
About Jittor:
The following content is from the official website of Jittor: https://cg.cs.tsinghua.edu.cn/jittor/:
Jittor: A deep learning framework based entirely on just-in-time compilation, using innovative meta-operators and a unified computation graph.
Meta-operators are as easy to use as Numpy and surpass Numpy in implementing more complex and efficient operations.
Installation Guide for Jittor:
Visit https://cg.cs.tsinghua.edu.cn/jittor/download/
Choose the corresponding installation steps based on your system.
How to Use the Project:
In PyCharm, simply run CGAN.py (ensure that Jittor is installed and compatible Python version is available, see "Installation Guide for Jittor").
数据集
CGAN_jittor
Overview:
About Jittor:
Installation Guide for Jittor:
How to Use the Project: