Face-to-Speech

generate speech from text & face image

Hi folks!👋🏻 This is The Prompt & we have some news: We polished our logo and the email format. Plus we have a new section: AI prompts, where we’ll feature interesting prompt hacks and results for you.

Hope you like it! Let’s see what’s new 👇🏻

HOT PICK

Generate speech from images and text

Can you imagine the voice of someone when you look at their face?

Well, this model can.

Face-TTS generates speaking styles and voices learned from facial characteristics.

Although the speech results still sound “artificial”, this is the first time that face images are used as a condition to train a Text-to-Speech (TTS) model.

You can see some of the results here.

INTERESTING TECH

Microsoft’s new model can understand images

KOSMOS-1 is the newest multi-modal model by Microsoft that can perceive general modalities, follow instructions, and perform in-context learning.

This means that the model is very good at many different tasks, like understanding language, talking to people, describing pictures, and answering questions about what it sees.

And it can do all of this with less training data, and/or fine-tuning.

The team also introduced a dataset of the Raven IQ test, which diagnoses the nonverbal reasoning capability of this model - or basically, it measures its IQ.

WHAT ELSE IS GOING ON

💬  Snapchat to launch its own chatbot (powered by ChatGPT). The big idea is that in addition to talking to our friends and family every day, we’re going to talk to AI every day.

👀  Elon Musk is recruiting AI researchers to create a ChatGPT rival. Musk actually co-founded OpenAI but has cut ties because of concerns about a conflict of interest with Tesla AI.

🏋🏻‍♀️ AI-designed COVID-19 drug starts clinical trials. They leveraged AI to find patterns and to develop a molecule that will become the basis of the drug.

RESOURCES

The best resources we came across lately that will help you become better at writing prompts & building AI apps.

👋🏻  OpenAI's Foundry leaked pricing says a lot – if you know how to read it [Insightful article about the AI in 2023]

TOOLBOX

The latest AI tools to use or get inspiration from.

  • Jailbreak Chat: A collection of ChatGPT jailbreaks

  • PyqAI: Build AI features with one simple API call

  • Mendable.AI: Chat-Powered Search Trained on Your Docs

  • Figma.AI: Integrate GPT Chat into your Figma workflow

  • editGPT: Easily edit and track changes to your content in chatGPT

PROMPT OF THE DAY

TOOL

Midjourney

PROMPT

💬 vertical farm 🍄🏰✨💫🥀☁️ --ar 4:5

RESULT

LATEST PAPERS

  • ChatGPT: A Meta-Analysis after 2.5 Months

  • SpikeGPT: Generative Pre-trained Language Model with Spiking Neural Networks

  • ELITE: Encoding Visual Concepts into Textual Embeddings for Customized Text-to-Image Generation

  • Internet Explorer: Targeted Representation Learning on the Open Web

  • Language Is Not All You Need: Aligning Perception with Language Models

  • Imaginary Voice: Face-styled Diffusion Model for Text-to-Speech

What'd you think of today's edition?

Login or Subscribe to participate in polls.