The field of AI has been in a state of rapid evolution. This article summarizes my learning, practice and observation articles in the past four months: the content covers AI large models and technical applications, technical details and tutorials, as well as artificial intelligence applications and development .
Artificial Intelligence Application and Development
What can AI large model LLM do to help enterprises?
AI Product List·Top 100 Domestic and Top 20 Overseas (August 2023)
Foreign reports indicate that 90% of AI product companies have achieved profitability, but interviews with domestic large models and AIGC said that this is too high.
The direction of development of large model applications｜The rise of agent and its future (Part 2)
The direction of development of large model applications｜The rise of agent and its future (Part 1)
Andrew Ng: Opportunities for AI
The winner of the 100-model war will eventually be the open source model | Near Craftsman
What is the "intelligence emergence" of AI, and why understanding it is of great value to entrepreneurs, practitioners, and ordinary people
8.23 Notes on China’s Big Model “Top Group Chat”
Is artificial intelligence safe? OpenAI is "aligning" large models with humans - ensuring ChatGPT is smarter than humans while still following human intentions
Will OpenAI go bankrupt by the end of 2024? Foreign media also love headline-making
Ant's large model is exposed, AI+ finance enters the "big model" era
Atom Capital: Big models are implemented in the financial field. It’s not easy to say that I love you
AIGC industry review
Large model dynamics and experience
OpenAI's new breakthrough: DALL·E 3 and ChatGPT collaborate to create comic strips with simple prompts
Beyond ChatGPT-4, Google’s multi-modal large model Gemini combined with AlphaGo technology has been tested on a small scale
What are the millions of ChatGPT users doing with it?
ChatGPT's "custom" function is open to free users. If the question information is insufficient, it will ask back to obtain the necessary information.
The world's largest open source translation model! Produced by Meta, supports 100 voices and languages!
The first batch of 8 large models in China are fully open! Large model applications enter a new stage
OpenAI releases ChatGPT fine-tuning interface! Domestic manufacturers are under great pressure! ｜Attached with detailed fine-tuning operation guide
ChatGLM update: LongBench—a dataset for evaluating long text understanding ability, supporting ChatGLM2-6B-32K with 32k contexts
Google Bard late night update: supports Chinese, voice input/broadcast, code export, conversation sharing
Bard! Google’s strongest counterattack against ChatGPT, Qing Mimi now supports Chinese!
OpenAI launches new product: using ChatGPT for content review
The Chinese large model Chinese-LLaMA-Alpaca-2 is open source and can be used commercially
Another top-notch large model open source is commercially available! Meta (Facebook)'s Llama 2 shakes up the big model melee landscape
ChatGPT update: Big killer! Function call demonstration
ChatLaw: Intelligent legal assistant from Peking University team, domestic large model successfully applies inclusive legal services
Everyone can make blockbusters! The movie effects produced entirely by AI explode!
Create a sliding lens effect like watching a blockbuster movie. Just a month ago, the AI drawing artifact Midjourney has been updated again.
Local version of ChatGPT; OpenAI launches Code Interperter
Performance increased by 571%, 32K ultra-long context, inference speed increased by 42%, commercial use allowed, domestic open source large model launched the second generation ChatGLM2-6B
ChatGLM2-6B first experience
Technical reports and analysis
Long text of 10,000 words - A review of large language model instruction tuning
Tongyi Qianwen-Qwen technical report details sharing
Reversing the Curse｜The Biggest BUG of Large Models! A=B Can’t the large model LLM trained learn B=A? Domestic models say yes
How to automatically identify high-quality instruction data from data sets - the use of IFD indicators
Technical science popularization and interpretation: hard-core interpretation of the ChatGPT large model! (1) Family history from GPT-1 to ChatGPT
Technical science popularization and interpretation: hard-core interpretation of the ChatGPT large model! (2) GPT4’s multi-modal emergence capability - close to key human characteristics
Technical science popularization and interpretation: hard-core interpretation of the ChatGPT large model! (3) Multi-modal architecture of GPT-4
Meta AI | Instruction back-translation: How to mine high-quality large model training data from a large number of unlabeled documents?
Revealing how WeChat trains large models: low-key WeLM｜The official website was last updated a year ago
Better than ReACT, it allows large models to learn the groundbreaking experience-based learning strategy ExpeL by "analogy by analogy" while solving problems.
REACT: Collaborating reasoning and action in language models, enabling them to solve a variety of linguistic reasoning and decision-making tasks.
Research on hallucinations of large language models | Alleviating and avoiding large model LLM hallucinations (2)
Siren Song in the Ocean of Artificial Intelligence: A Review of Hallucination Research in Large Language Model LLM (1)
Hello, I am Baichuan Big Model｜The secret of Baichuan2, which is open source and free for commercial use in China
Large model distributed training parallel technology (3) - pipeline parallelism
How to conduct fine-tuning experiments on large models and record a fine-tuning experiment process on a large model based on ChatGLM-6B.
Large model localization adaptation 2-using ChatGLM-6B for model inference based on Shengteng 910
From memory wall to efficient computing｜How to improve the computing speed of large models in inference: MQA and GQA optimization strategies
A complete guide to getting started with LLM large model technology in the ChatGPT era! All top conference papers and practical information in one place
How MM-Vet’s multimodal evaluation criterion evaluates the performance of large multimodal models (LMM) on complex tasks
Detailed interpretation of LLama2 | How does LLama2, the light of Meta open source, catch up with ChatGPT?
The ChatGLM team released the AI Agent capability evaluation tool AgentBench: GPT-4 is the best, but the open source model performs very poorly!
In-depth analysis of Llama 2 resource summary: not to be missed
M3E is probably the most powerful open source Chinese embedding model
Claude 2 explains the technical secrets of ChatGPT 4: details: number of parameters, architecture, infrastructure, training data set, cost
Why is it that if you say "You are an expert in such and such a field" to large language models such as ChatGPT and ChatGLM, its answer will be much more effective? (three)
Why is it that if you say "You are an expert in such and such a field" to large language models such as ChatGPT and ChatGLM, its answer will be much more effective? (two)
Why is it that if you say "You are an expert in such and such a field" to large language models such as ChatGPT and ChatGLM, its answer will be much more effective? (one)
The official Android version of ChatGPT is officially released! Two detailed download and installation tutorials
Large model application
Jupyter AI: Generate code, correct errors through chat, and support various large models
Hayo AI: A must-have medicine for home travel for domestic friends, integrating ChatGPT and drawing Stable diffusion
ChatALL: The amazing AI robot that finds the best answers!
Installation tutorial
Local installation of Stable Diffusion tutorial Mac version
Use Alibaba Cloud's free serverless function computing product resources to install Stable Diffusion
No hardware resources? Use Colab to build your own Stable Diffiusion online model for free! Nanny level tutorial
Quickly deploy the Tsinghua large model ChatGLM2-6B, and handle HuggingFace Space with one click
Development practices and open source projects
Recommend FinGLM, an open source large model project for financial analysis
3 amazing GitHub open source projects were born!
Recommend three open source projects to build proprietary knowledge base + large model intelligent assistant
Play with PDF chatbot in 5 minutes! Super simple Langchain+ChatGPT implementation strategy
The key to unlocking artificial intelligence project development: Detailed explanation and advanced learning of Python basic library
Quickly connect the intelligent chat robot to the domestic ChatGLM open source model in 1 minute
ChatGPT development practice｜A small tool to translate English subtitles into Chinese bilingualism
Tutorial | Use free GPU resources to build an exclusive knowledge base ChatGLM2-6B + LangChain
Project implementation practice｜Building an exclusive knowledge base based on ChatGLM2-6B + LangChain is initially completed
Project implementation practice｜The domestic large model ChatGLM2-6B was successfully deployed on Alibaba Cloud
Implementation function: ChatGPT WeChat assistant can automatically search network information to answer questions
ChatGPT WeChat assistant is online! Questions and answers are more convenient, and functions are being continuously upgraded.
Customized needs | How high is the computing power cost of a large artificial intelligence model application?
PromptPrompt
How to avoid the nonsense of big language models such as ChatGPT and ChatGLM
How to use ChatGPT? Prompts for 12 scenarios & universal conversation templates & four questioning skills
Why does it seem that the prompt word Prompt you are using in ChatGPT does not work as expected?
A guide to how to use Prompt! Six strategies to make ChatGPT smarter (Part 2)
A guide to how to use Prompt! Six Strategies to Make ChatGPT Smarter (Part 2)
Ignite creative inspiration: Prompt practical guide revealed! Six strategies to make ChatGPT smarter (Part 1)
Practical Teaching Prompt Prompt Word Practice: How to use ChatGPT to guide college entrance examination Chinese composition writing
Large model security
Prompt attack attacks large models again, hypnotized ChatGPT may leak important information - hidden risks of large models
Open ChatGPT "dangerous speech" with one click! Research finds: AI chatbot actually has a “big bug” that cannot be fixed at present
The large model of AI artificial intelligence is lost! ChatGPT, BARD, BING, and Claude were successively attacked by "prompt attacks"!
Confirmed: GPT cannot provide valid Windows 11 keys
Two tips to make the most of ChatGPT
Improve efficiency and use ChatGPT to easily write daily and weekly reports
Are you sure you don't want the ChatGPT prompt word Prompt sold for 2w?
Embrace the future and learn AI skills! Follow me and receive free AI learning resources.