GoJun AI Daily

English

简体中文

English

简体中文

Appearance

Today's selected cutting-edge AI information, welcome to read 👇

🗣️ News reveals that GPT-4o Advanced Voice Mode will be fully open on the 24th!

🎨 DrawingSpinUp generates high-quality 3D animations from hand-drawn images without training!

📄 Open-source and free document scanning tool OpenScanner, fast, easy to use, and powerful!

🗺️ Real-Address-Generator generates real random addresses based on Cloudflare Workers!

📚 Andrew Ng recommends the book "Hands-On Large Language Models," covering practical LLM essentials with open-source code!

Cutting-edge News ​

1. News reveals that GPT-4o Advanced Voice Mode will be fully open on the 24th.

Anticipation has been building since the Spring launch event. Hopefully, it will arrive on time, or else it risks being overtaken.

Source: https://x.com/testingcatalog/status/1836523671554527472

Image

Cutting-edge Technology ​

1. A tool that can generate animated 3D models from single-character hand-drawn images: DrawingSpinUp.

No training required; it can be used directly to transform a single hand-drawn character into a high-quality 3D animation while maintaining the style and details of the drawing.

Detailed introduction: https://lordliang.github.io/DrawingSpinUp/

GitHub: https://github.com/LordLiang/DrawingSpinUp

The code is open-source and deployable. This technology allows children's hand-drawn pictures to come to life, which is very fun.

image

Open-source Projects ​

1. A fast, easy-to-use, and open-source free document scanning tool: OpenScanner.

GitHub: https://github.com/pencilresearch/OpenScanner

Key features include:

  • Automatic text recognition to generate page transcripts;
  • Automatic document naming;
  • Tagging scans by current geographic location;
  • Adding signatures imported from other devices;
  • Native support for Vision Pro;
  • Saving scanned documents in the Open Scanner library;
  • Editing and reordering scanned document pages;
  • iCloud sync to all devices;
  • Annotating and editing scanned documents;
  • Using Apple Pencil to mark scans on iPad;
  • Easy sharing and exporting of scanned documents as PDFs;
  • 100% native iOS app, no reliance on third-party accounts;
  • Open-source and free, no ads or subscription required.

image-20240920150843883

2. A real address generator based on Cloudflare Workers: Real-Address-Generator.

It generates real random addresses on maps of different countries, including names, gender, phone numbers, addresses, and other information.

GitHub: https://github.com/Adonis142857/Real-Address-Generator

img

Learning Resources ​

1. Sharing a book recommended by Andrew Ng: "Hands-On Large Language Models."

The book has twelve chapters, covering fundamental knowledge of language models, text classification, prompt engineering, semantic search, multimodal language models, and how to create and fine-tune text embedding models.

All the case study code from each chapter has been open-sourced on GitHub and can be run directly through Google Colab.

GitHub: https://github.com/handsOnLLM/Hands-On-Large-Language-Models

Additionally, the book is enriched with over 250 customized illustrations and supported by runnable code, timelines, and references to key papers to help readers better understand and use large language models.

img