Spend 1 minute each day to get curated cutting-edge AI information.
The content includes but is not limited to cutting-edge AI news, AI tools, AI art, open-source projects, and learning tutorials, etc.
Follow the AI Daily to keep up with the AI trend, and I hope it will be helpful to you. For important information, separate posts will be made for detailed introductions.
Below is the latest AI information for August 10th.
Cutting-Edge News
1. Alibaba's audio model Qwen2-Audio releases an online demo.
You can interact with the model through voice input, and the model will analyze and interpret various types of audio information based on voice commands.
Online demo: https://huggingface.co/spaces/Qwen/Qwen2-Audio-Instruct-Demo
Currently, it supports 8 languages including Chinese, Cantonese, English, French, Japanese, etc. I was quite surprised to see that Cantonese is supported.
AI Art
1. Google's image generation model Imagen 3 is now open to everyone!
It supports partial re-drawing, allowing you to edit images with brushes and prompts.
Usage link: https://aitestkitchen.withgoogle.com/zh/tools/image-fx
Note that it is temporarily unavailable in some regions; I have personally confirmed that it works in Japan.
2. The FLUX product for image generation with ControlNet is now available for online experience!
Experience link: https://huggingface.co/spaces/DamarJati/FLUX.1-DEV-Canny
Currently, only the Canny mode of ControlNet is supported.
Learning Tools
1. An interactive learning tool for Transformer visualization: Transformer Explainer.
Using GPT-2 as an example, you can input text to observe in real-time how each component (embedding layer, self-attention mechanism, MLP, etc.) processes the text and predicts the next token.
GitHub: https://github.com/poloclub/transformer-explainer
Learning link: https://poloclub.github.io/transformer-explainer/
This tool, which combines visualization and interaction to help users understand the internal components of Transformers, is especially suitable for teaching. It's very illustrative, so if you're interested, give it a try.
