Pager 2 - AI blog - Learn AI from scratch

[AI SenseVoice] Alibaba Open-Sources Speech Foundation Model with Automatic Speech Recognition, Language Identification, Speech Emotion Recognition, and Audio Event Detection

Alibaba open-sources the SenseVoice speech foundation model, featuring high-accuracy multilingual speech recognition, emotion recognition, and audio event detection capabilities. Supports over 50 languages, outperforming the Whisper model.

[AI Cog] Want to Run an AI Business but Lack a GPU? Cant Set Up the Environment? Use Cog to Easily Deploy Your Business to the Cloud

Explore how to deploy your AI business to the cloud using Cog, achieving serverless deployment without a GPU.

[AI CrewAI] Be the Boss, Form an AI Team, and Collaborate with AI Agents to Complete Tasks

Learn how to achieve AI agent collaboration with CrewAI to boost productivity and efficiency. CrewAI provides a solid foundation for building intelligent assistant platforms, automated customer service suites, and multi-agent research teams.

[AI Programmer Devyan] Finally Found Someone to Write Code for Me

Learn how Devyan, an AI-driven programming assistant, simplifies and optimizes the software development process by coordinating multiple roles.

[AI Fabric] This Is How Prompts Should Be Used

Fabric is an open-source framework designed to enhance human capabilities through artificial intelligence, addressing the challenges of AI integration. By breaking down problems and applying AI solutions, Fabric helps users tackle daily challenges, achieving a perfect blend of technology and humanity.

[AI MoA] So This is MoA (Mixture-of-Agents)

Introducing a project, Mixture-of-Agents (MoA). MoA is an innovative approach that leverages the collective advantages of multiple LLMs to enhance performance, achieving state-of-the-art results. By adopting a hierarchical architecture with multiple LLM agents at each layer, MoA uses only open-source models and scores 65.1% on AlpacaEval 2.0, significantly outperforming GPT-4 Omni's 57.5%!

AI Kuaishou LivePortrait

[AI Kuaishou LivePortrait] Leading the New Era of Efficient Portrait Animation

Kuaishou has launched LivePortrait, featuring efficient portrait animation with stitching and redirection control. Download the code, prepare the environment, download pre-trained weights, and start inference.

AI

[AI Kimi Extension] Kimi Browser Plugin: Instantly Enhance Your Browsing Experience with AI

The Kimi browser plugin provides an easy installation and usage guide, helping users quickly install the plugin and take advantage of its powerful features, such as text highlighting for questions, article summarization, and inspiration for writing.

[AI Perplexica] Installation Guide: Easily Deploy AI-Powered Open Source Search Engine

A detailed guide on how to deploy the AI-powered open-source search engine Perplexica using Docker in your local environment, allowing you to quickly get started and experience its powerful features.

[AI Kimi] Context Caching Public Beta, Reducing Long Text Model Costs by 90%

Kimi's Context Caching technology is now in public beta. This technology pre-stores data to significantly reduce computing costs and latency, suitable for long text models, helping to save up to 90% of costs and reduce first token latency by 83%.