GoJun AI Daily

English

简体中文

English

简体中文

Appearance

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

💻 Alibaba releases new generation programming model Qwen2.5-Coder, offering 4 parameter scales, 32B version performance comparable to GPT-4, supporting Cursor integration and visual Artifacts.

🤖 OpenHands multi-agent platform turns AI into programmers, capable of executing code modifications, command running, API calls, and other development tasks, providing Docker deployment.

🎙️ Voice-Pro open-source tool integrates speech-to-text, translation, and TTS functions, supporting 100+ languages, providing noise reduction and subtitle export features.

Latest News ​

1. Alibaba releases latest programming model: Qwen2.5-Coder full series.

It comes with four different parameters: 0.5B, 3B, 14B, and 32B, all of which can be integrated with Cursor and feature the ability to generate visual Artifacts panels similar to Claude.

Detailed introduction: https://qwenlm.github.io/zh/blog/qwen2.5-coder-family/

Model download: https://huggingface.co/collections/Qwen/qwen25-coder-66eaa22e6f99801bf65b0c2f

The 32B parameter model rivals GPT-4 in multiple benchmark test scores, demonstrating powerful and comprehensive coding and mathematical capabilities.

img

Cutting-edge Technology ​

1. An AI-driven software development multi-agent platform: OpenHands, claiming to replace "programmers".

Composed of multiple agents forming an "AI programmer," capable of executing various real-world development tasks, including code modification, command execution, web browsing, API calls, and even copying code snippets from StackOverflow.

GitHub: https://github.com/All-Hands-AI/OpenHands

The project provides detailed installation and deployment tutorials, with quick start via Docker. It's somewhat similar to an open-source version of Cursor or Bolt.

App screenshot

Open Source Projects ​

1. An open-source tool integrating transcription, translation, and text-to-speech: Voice-Pro.

Provides a clean and intuitive visual interface, one-click installation, supports real-time transcription and translation, as well as batch processing mode.

GitHub: https://github.com/abus-aikorea/voice-pro

Main features:

  • Provides integrated environment for YouTube downloader, noise removal, subtitles, translation, and TTS
  • Supports speech recognition and text-to-speech in over 100 languages
  • Supports speech-to-text using Whisper, Faster-Whisper, etc.
  • Supports TTS voice speed, volume, and pitch adjustment
  • Provides word-level highlighting and noise reduction features
  • Supports export of multiple subtitle file formats, such as ass, srt, tmp, ssa, etc.
  • Supports output of multiple audio formats, such as wav, flac, mp3, etc.

img