Deep learning has made OpenAI shine, releasing a remarkable model, and they’ve dropped something amazing as well. When we talk about image models, DALL-E is considered state-of-the-art as of today. DALL-E generates very large images; the only problem is that this is offered by OpenAI, which is actually not open but closed-source. After releasing the open-source model with DALL-E, the market capitalization in US dollars has dramatically shifted. They released another model in machine learning and also made it open-source. This means that what OpenAI used to offer DALL-E 3 for a fee is now available for free through ChatGPT. Let me show you how we can use it.
As you can see, I am now on the computer screen and I will type "Gens Pro" and here I will write "Hugging Face." Now, you can see Gens Pro available on Hugging Face. If I click on Spaces and go to the search spaces, I will find this Gens Pro space. I’ll share the link with you, but let me first click on "Drop Image" to upload an image where we see some people standing in New York. I will ask it, “What is this image?” and "What’s the technology behind it?" Now I will click on "Chart." Because this is a very hot model, many people are trying it out, and it will take some time to get a GPU. I will fast-forward so you can see what the results look like, but you will need to keep clicking if you're not getting a GPU. It will acquire a GPU and generate the results for you.
Now you can see that this image of FPS 2 has many details; it identified the city, and it even pulled from Google simply by typing in "New York." It provides details about the watermark and the individuals in the image. Notice how effectively it understood the image since it’s a substantial model. The exciting part here is that in the benchmarks with GNU and JPEG, it surpassed DALL-E 3. This open-source capability is a significant development because you can run it locally. Today, we have very powerful GPUs, and in a couple of years, we may see GPUs that are two to three times more powerful and affordable. This will allow such models to be run locally, making it entirely feasible.
Now, I want to tel
Post a Comment
0Comments