In the past week, the AI world has undergone earth-shaking changes. Among them, the most well-known one is that OpenAI officially launched the ChatGPT iOS client , so that everyone can chat with ChatGPT on their mobile phones more conveniently.
In addition, Stability AI, the parent company of Stable Diffusion, has also open-sourced an AI image generation platform StableStudio on GitHub . In the future, it will successively access multiple image generation large models, so that everyone can deploy an AI painting platform similar to Midjourney for free.
In recent days, many new projects have sprung up on GitHub. In addition to the open source projects that I tweeted earlier, some good tools and solutions have also appeared recently.
Below, I will share with you three open source projects that have been active on GitHub recently, have commercial application potential and are extremely practical.
Without further ado, let's get down to business.
Open source ChatGPT code interpreter plugin
On the day when the ChatGPT plug-in system was officially released, in addition to Web Browsing, there was another powerful plug-in that was often overlooked.
That is the Code interpreter (code interpreter) plugin.
You can talk to ChatGPT, let the plug-in generate various data visualization charts for you, automatically analyze the chart data, and edit the size and tone of the picture online with instructions.
Unfortunately, although ChatGPT has opened various plug-in permissions for all Plus users, this plug-in is still only available to a small number of users.
A foreign developer applied for the plug-in permission many times and waited for no results. He couldn't hold back and made a simplified solution: GPT Code UI .
This project can take advantage of the large language model capabilities to automatically generate and execute code. In addition, it also supports file upload and download, context understanding, and optional GPT-3.5 and GPT-4 models.
GitHub：https://github.com/ricklamers/gpt-code-ui/
You can use Python's pip installation package tool to quickly install it locally:
$ pip install gpt-code-ui
$ gptcode
You can directly upload a CSV data table file, and through a dialogue with ChatGPT, let it automatically add, delete, modify and check the file, and output a file download link after automatic saving.
Or you can throw it a txt text file, let it scan the content of the file, and combine specific data to generate an analysis report for you.
Build your second brain with AI
After the emergence of ChatGPT, its powerful document summary and retrieval capabilities are always shocking.
You can let ChatGPT learn and evolve by feeding it a document, enterprise knowledge base, and tutorials, and transform it into an enterprise intelligent customer service, training instructor, financial advisor, etc.
Today I would like to recommend an open source AI knowledge base solution: Quivr .
The project supports uploading text, images, videos, code snippets, PPT, Excel data tables and other content directly to the cloud database, and quickly realizes information retrieval and question answering through large language models such as GPT-4.
, duration 01:19
GitHub：https://github.com/StanGirard/quivr
The project supports large language models such as GPT-3.5, GPT-4, and Claude, and can be deployed by itself to ensure privacy and security.
After the emergence of AI, we need to change the way we learn and think in the past. Since ChatGPT is so powerful, we should make full use of it and hand over some time-consuming tasks to it.
Train AI as your second brain, let it help you complete various information intelligent retrieval and question answering tasks, and further improve productivity.
One-stop AI chatbot
It has to be said that this year, OpenAI has pushed the technological wave of large language models to the apex on its own.
Since the birth of GPT-4, multiple big languages have sprung up like mushrooms after rain, which is really dazzling.
In order to allow users to fully experience the capabilities of each large language model, quickly realize the model switching operation.
A domestic developer officially open-sourced a cross-platform client called ChatAll on GitHub.
Allows you to have conversations with multiple AI robots at the same time, and become the king of the AI world
GitHub：https://github.com/sunner/ChatALL/
The tool has the following features:
-
Quick prompt mode: multiple prompts can be sent continuously;
-
Store chat records locally;
-
Automatically highlight reply content;
-
Keep ChatGPT offline;
-
disable/enable arbitrary bots;
-
Multi-language support, multi-view switching.
Support more than a dozen mainstream AI chatbots such as ChatGPT, Bing, Bard, Claude, ChatGLM, etc.
Since ChatAll is a cross-platform chat client, you can quickly download its corresponding Windows, macOS, and Linux applications from the releases page of the GitHub project.
Well, the above is the ChatGPT open source project recommended to everyone this time.