AI based Interior Designer built on Google AI Studio
Welcome!
You must be living under a rock if you haven’t seen the latest Generative AI updates from Google. In the past few weeks Google has gone ahead and released their most powerful model Gemini Pro additionally, it has renamed MakerSuite as Google AI Studio and has increased the capabilities of Makersuite and the PALM API.
Amongst these updates, one particular one stood out, which was the Gemini Pro vision model that is available inside of Google AI Studio. This model has the capabilities to understand images as prompts and give back a text response.
You can check out all the updates and advancements Google has made in Generative AI by checking out the blog below.
In this article, we will be exploring the capabilities of the Gemini Pro Vision model that can accept photos as prompts and reply based on the question asked in the prompt.
Once we test the capabilities of the model, we will be building an AI interior Designer built using the Gemini Pro Vision model on Google AI Studio.
This article and all the articles that will be coming soon, will not be covering the basics of Google AI Studio(formerlly known as MakerSuite). If you want to learn from the basics check out the YouTube playlist given below that gives you a walkthrough for Google Bard, MakerSuite and even the PaLM API (known as Gemini API).
Launch Google AI Studio
Now that we have the basics cleared, lets go ahead and dive deep into Google AI Studio’s Gemini Pro Vision Model. To launch Google AI Studio, you can either make a quick google search or click here.
On opening the Google AI Studio, you will be directed to your prompt library that looks something like this.
Currently, I have multiple projects built on it but you can go ahead and create a new Freeform prompt. To create a new Freeform prompt, go ahead and click on Create new and click on FreeForm Prompt.
To know more about the other types of prompts available, check out blog link given below.
On creating a new Freeform prompt, you will see something like this.
Currently this particular project is working on the Gemini Pro model that only accepts text as prompts. To integrate the power of Gemini Pro Vision model, go ahead and change the model on the right side of your screen.
If you have previously tuned models, you will also be seeing them here as an option, but go ahead and click on Gemini Pro Vision.
Once you select the Gemini pro Vision model, you will see something like this.
This now indicates you can add images to your prompts and ask questions on the images that you pass.
To test the capabilities of this particular model we are going to be testing the example prompts that have been provided.
We are going to be asking it to suggest us meals based on some dishes. On successful execution you will see that it has generated a meal plan for us.
AI based Interior Designer
Now that we have understood the basics of what this model can do, lets go ahead and build over own AI powered interior designer using the Gemini Pro Vision model on Google AI Studio.
The plan is to first pass a photo of my living room, followed by providing it with two decoration pieces that I intend to attach to the wall. We are going to be testing if its able to differentiate between the two items and suggest us which item will look better on the wall.
I have first attached an image of my living room followed by the first item that I want to attach on the wall. Under each image that I have attached, I have specified a text that describes the images.
After specifying the living room photo and the first item, I have attached the second item that I want to apply on the wall.
Here I have successfully added all the details to the prompt and we can go ahead and click on Run and unveil the power of the Gemini Pro Vision model.
On clicking Run, in a few seconds you will see that the model has responded to your prompt and given you a valuable suggestion to the question specified in the prompt.
Here you can see on successful execution, we have passed images in the prompt and got a desired output as well.
Conclusion
This is a very basic but interesting example to actually understand the capabilities of this model. If you want to know more about Gemini Pro, you can check out the link given below.
If you want a video version of this blog, you can check out the YouTube tutorial given below.
Feel free to reach out if you have any issues/feedback at aryanirani123@gmail.com.
