Getting Started with Mistral: An Introduction The open-source Mixtral 8x7B model launched by Mistral adopts a “Mixture of Experts” (MoE) architecture. Unlike traditional Transformers, the MoE model incorporates multiple expert feedforward networks (this model has 8), and during inference, a gating network is responsible for selecting two experts to work. This setup allows MoE to … Read more
Machine Heart Column Machine Heart Editorial Team Fusion of multiple heterogeneous large language models, Sun Yat-sen University and Tencent AI Lab introduce FuseLLM With the success of large language models like LLaMA and Mistral, many major companies and startups have created their own large language models. However, the cost of training new large language models … Read more
Machine Heart Reports Editor: Demon King This extension utilizes the pre-trained model CodeBERT, which can handle programming languages and natural language, to quickly generate Python docstrings. Visual Studio Code (VS Code) is a free code editor developed by Microsoft that supports operating systems such as Windows, Linux, and macOS. It supports testing and has built-in … Read more
Report by Machine Heart Editor: Zhang Qian This nearly one hundred page review outlines the evolution of pre-trained foundation models, showing us how ChatGPT has gradually achieved success. All successes have a traceable path, and ChatGPT is no exception. Recently, Turing Award winner Yann LeCun was trending due to his overly harsh evaluation of ChatGPT. … Read more
[Introduction] The emergence of BERT has revolutionized the model architecture paradigm in many natural language processing tasks. As a representative of pre-trained language models (PLM), BERT has refreshed leaderboards in multiple tasks, attracting significant attention from both academia and industry. Stanford University’s classic natural language processing course, CS224N, invited the first author of BERT, Google … Read more
Pinecone + LangChain: Building an Efficient AI Retrieval System Recently, while experimenting with AI retrieval systems, I found that Pinecone and LangChain are a match made in heaven. Pinecone is a powerful vector database, and LangChain is a flexible framework. Combining the two allows you to easily build an efficient AI retrieval system. Today, I … Read more
On July 7, Alibaba Cloud announced at the 2023 World Artificial Intelligence Conference that the AI painting creation model Tongyi Wanshi has begun targeted testing. Tongyi Wanshi is the third product in Alibaba Cloud’s “Tongyi” model series, following Tongyi Qianwen and Tongyi Tingwu, which feature text Q&A and speech-text processing capabilities respectively. The first three … Read more
▌A Little Story About Demand Seeing the Scene When we publish articles on public accounts, we need some cover images. Is there an AI tool that can quickly generate images that can visually express the content of the article? Identifying the Problem Some of the cover images for the public account articles are edited by … Read more
Hello everyone, I am Dong Ge! A few days ago, Tongyi officially launched theTongyi Wanshang 2.1 series, significantly enhancing its image generation and video generation capabilities, with great improvements in UI optimization, Chinese character generation, complex camera movements, generation stability, and clarity. The New Version of Tongyi Wanshang Has Greatly Improved Capabilities Using the recommended … Read more
The date is January 9, 2025. I came across an article on a public account about the upgrade of Tongyi Wanxiang, praising its camera work and quality as professional-grade. Thus, I had the idea to experience and compare various domestic AI video generation tools. Note that I am not a professional user and not skilled … Read more