Today's selected cutting-edge AI information, welcome to read 👇
🎬 Kuaishou open-sources Pyramid Flow video generation model: Supports text-to-video and image-to-video, with excellent results.
📱 Notify Me: Open-source Android application, captures phone calls and SMS and forwards them, solving multi-device information reception issues.
🎙️ Podcastfy: Open-source tool that can convert various content formats into conversational podcasts, similar to NotebookLM.
Cutting-edge Technology
1. Kuaishou open-sources a new video generation model: Pyramid Flow.
Supports text-to-video and image-to-video generation, capable of producing 768p, 10-second videos at 24 FPS, with a parameter count of 2B.
Detailed introduction: https://pyramid-flow.github.io/
From the demonstration videos, the results look very good. Additionally, someone has open-sourced its ComfyUI workflow for those interested in trying it out.
ComfyUI workflow: https://github.com/kijai/ComfyUI-PyramidFlowWrapper
Open Source Projects
1. Recommending an open-source free Android application on GitHub: Notify Me.
This application can capture incoming calls and SMS on your phone and forward the data to a Bark server or email, solving the problem of not receiving important information in a timely manner across multiple devices.
GitHub: https://github.com/jinweijie/notify-me
Features:
- Smart message capture: Automatically captures incoming calls and SMS on Android devices;
- Multi-channel forwarding: Simultaneously forwards information to Bark server and email, ensuring timely reception with dual protection;
- Cross-device notifications: Receive notifications on any device through the Bark app;
- Privacy protection: Supports setting up a private Bark server for data security and privacy.
2. Discovered an open-source tool that can replace NotebookLM: Podcastfy.
Input video, PDF, papers, websites, articles, and other content, and convert them into conversational podcasts with one click, similar to NotebookLM.
GitHub: https://github.com/souzatharsis/podcastfy-demo
Those interested can try it out on HuggingFace. An API Key is required, and it's recommended to use the Gemini API.
Online experience: https://huggingface.co/spaces/thatupiso/Podcastfy.ai_demo