Spend just 1 minute daily to get comprehensive updates on AI technology developments, industry trends, and market insights.
The content includes but is not limited to cutting-edge AI news, AI tools, AI painting, open-source projects, and learning tutorials.
Stay tuned to AI Daily for the latest in AI trends. For important information, detailed posts will be made separately.
Here is the latest AI information for July 19.
Cutting-edge News
1. OpenAI releases GPT-4o mini, replacing GPT-3.5!
This is truly impressive, achieving the best balance between window size, cost, speed, and performance while enhancing overall intelligence levels.
- Window: Features a 128K context window with a maximum output of 16K (currently the largest), ideal for RAG;
- Price: 60% cheaper than GPT-3.5 Turbo;
- $0.15 per million input tokens;
- $0.6 per million output tokens;
- Performance: Scores 82% on MMLU, preferred in chat preference tests over GPT-4, and has math reasoning and code abilities close to 4o.
Detailed introduction: https://openai.com/index/gpt-4o-mini-advancing-cost-efficient-intelligence/
Available for all users, including free users, on the official website or via API.
2. Mistral releases a 12B model, Mistral NeMo.
With a 128K context length, it performs best in reasoning, knowledge, and code accuracy among models of the same size.
Model download: https://huggingface.co/mistralai/Mistral-Nemo-Base-2407
Cutting-edge Technology
1. Another AI virtual try-on project, IMAGDressing-v1.
Achieves realistic virtual try-on effects using latent diffusion models, allowing free control of clothing, face selection, poses, and scenes, significantly enhancing the online shopping experience.
Supports integration with ControlNet and IP-Adapter to enhance the diversity and controllability of generated images.
GitHub: https://github.com/muzishen/IMAGDressing
Online experience: https://sf.dictdoc.site/
Open-source Projects
1. An open-source free visual web scraping platform, Kspider.
Allows configuring web scrapers through flowcharts, making data scraping possible without writing code, extremely user-friendly for beginners.
GitHub: https://github.com/kkangert/kspider
Main features include:
- Supports xpath and css selectors;
- Supports automatic cookie management;
- Supports plugin extensions (custom executors, custom functions, etc.) and hot-swapping;
- Supports scraping dynamically rendered pages by JS;
- Supports synchronous and asynchronous execution;
- Supports custom JS script engine.
Provides a convenient management UI for online management and debugging tasks, viewing task logs, and visual debugging of scrapers, doubling the efficiency of our scraping work.