AutoNetGen

AutoNetGen (Auto Network Generator)

Let AI design AI, let big models help small models evolve, and use magic to create magic!

Empower Artificial Intelligence to sculpt its own kind, where colossal models gracefully usher the petite ones into evolution, weaving magic to conjure further enchantment!

AutoNetGen

Let AI design AI, let big models help small models evolve, and use magic to create magic!

• Project name : AutoNetGen (Auto Network Generator)

• Version : 0.6 beta

• Core Driver : GPT-4 Advanced Data Analysis

• Author : Yunzhongjiangshu

Function

• Existing features

• AI model design

• AI model optimization

• AI model deployment

• Code function verification

• Code Bug Repair Self

• To be completed

• Data preprocessing

• Training Verification

• Machine Learning Model Design

• Quantitative analysis of AI model

Installation and Run

• Continue AI model design: You can click on the link to continue model design

• To start over, please turn on the Advanced Data Analysis (formerly Code Interpreter) function of GPT-4 and copy and paste all contents in AutoNetGen.txt into the dialog box to run.

Prerequisites

️ Code execution is required, so the GPT-4 with Advanced Data Analysis enabled (Code Interpreter) function must be used. Currently this feature is only available to ChatGPT Plus users.

️ The code generated by ChatGPT may have unknown bugs. Please check carefully before using it in a production environment!

Get updates

This work is updated from time to time. Please visit the Github repository of this project to obtain the latest version.

Use preview

Initial environmental inspection

AI model design

AI model optimization

AI Model Deployment—Export ONNX

Complete conversation

Please refer to this complete conversation before use:

• AutoNetGen

Known issues

Due to the instability of the OpenAI Advanced Data Analysis environment state itself and the limitation of GPT's own capabilities, the following known problems exist (if you have a better solution, please contact the author):

1. The inspection results cannot be displayed correctly: the environment inspection code is executed successfully, and the execution results cannot be displayed correctly on the web page.

2. File cannot be downloaded: File download link is invalid

3. Environment status loss: environment change, previous execution results are lost

4. The downloaded code file content is incorrect: find the corresponding steps, click show work to view the original code, you can manually copy the code and save it

5. Resources exceed limit: Too large network leads to insufficient memory

Open source and cooperation

This game is an EmbraceAGI open source community project. * Welcome to join the EmbraceAGI open source community, and welcome friends from the EmbraceAGI community to walk through the waves in the AI ​​era!

Community Sister Project

• LangGPT structured prompt words, please click the link to visit the LangGPT project.

• Life Repeat Simulator

protocol

This project uses the CC BY-NC-SA 4.0 (Creative Commons-Attribution-Non-Commercial-Share 4.0 International in the same way) protocol.