Microsoft quietly released a system called Visual ChatGPT based on ChatGPT, which is a system that uses ChatGPT to generate drawings based on text prompts. The traditional "you say, I draw" systems like Stable Diffusion are well-known, so what makes V ChatGPT different from them?
Taking this opportunity, I decided to use the ChatGPT toolchain to interpret the newest member of the ChatGPT family.
Since we mentioned the ChatGPT toolchain, let's establish a rule: starting from this point, except for step-by-step instructions on tool usage, all text will come from the ChatGPT toolchain. First of all, the original ChatGPT cannot access the internet and its data only goes up until 2021, so it doesn't have any content about Visual ChatGPT. Therefore, we need additional tools to provide it with relevant information.
Fortunately, the Github page of Visual ChatGPT comes with their paper, which provides detailed explanations. So, we can use the ChatPDF tool developed by @xathis to read the paper.
First, let's ask it what V ChatGPT is. Here is the generated content for "Give me a summary about Visual ChatGPT":
It looks great, but what about students who don't understand English? We can ask it to summarize in Chinese.
Of course, you can also use the openai-translator developed by yetone.
However, both English and Chinese exceed the character limit for a tweet. Can we make it more concise? Yes, here is its answer:
Visual ChatGPT is a system that combines different visual models, allowing users to interact with AI systems through language and images. It supports complex visual tasks, including visual editing instructions that require collaboration between multiple AI models. The system has been validated through extensive experiments for its understanding and generation capabilities.
Let's ask something more complex: What is the difference between Visual ChatGPT and Stable Diffusion?
Please provide a detailed explanation of the workflow in Visual ChatGPT.
Further explain the Prompt Manager in Visual ChatGPT.
Give an example of using Visual ChatGPT.
How should I install and use Visual ChatGPT? The PDF doesn't provide any information.
That's all for this content. As you can see, with the help of automation, ChatGPT can replace a lot of manual work recommended by various technologies and tools.
My Twitter: