Post AI Academy

of the Week 😎

New neural network for voice-to-voice translation, Midjourney update, free Code Llama neural network for code generation, and generative AI in Paint!

Here's what stood out this week in the world of AI 👇

1️⃣ Another update has rolled out for Midjourney.

We've already written about it here. Now, MJ has its own version of the Generative Fill function and the ability to change parts of an image using the Inpainting function.

You can see a visual comparison of Generative Fill in Midjourney and Photoshop in the video attached to this post.

2️⃣ Think everyone forgot about Paint? Not so! Microsoft is training its creation and integrating generative AI into it.

According to a Windows Central report, Microsoft is exploring ways to implement artificial intelligence features into the app in Windows 11.

In the near future, the company plans to develop a neural network for Paint based on Bing Image Creator, allowing you to generate based on sketches and text.

3️⃣ A new neural network from Meta breaks down language barriers.

This week saw the release of the SeamlessM4T neural network.

It translates voice to voice, but in another language. Text translation is also available.

For now, only a demo version is available, but a full release is expected soon.

4️⃣ Not just text translations, but also code writing: another big release of the week from Meta is the programmer's assistant Code Llama. It's free for now.

By the way, many experts have already rated Llama higher than GitHub Copilot.

You can test it here.

What discoveries in the field of neural networks have you made this week? We'd love to hear in the comments 👇