Application Frameworks, Libraries, and APIs

  • DALL·E 2 is the OpenAI deep learning model for generating images from prompts, or natural language descriptions
  • Stable Diffusion is the open source AI image generator and model from Stability AI
  • Midjourney is an AI image generator accessible through prompting a Discord bot
  • Stanford Alpaca provides an instruction-following LLaMA model
  • Dalai provides an NPM for running LLaMA locally
  • whisper.cpp is OpenAI’s Whisper ASR model inferene in plain C/C++ without dependencies
  • llama.cpp is Meta’s LLaMA 7B-parameter model inference in plain C/C++ without dependencies
  • alpaca.cpp provides code to locally run a fine-tuned, instruction-following Stanford Alpaca model
  • MiniLLM is a minimal system for running modern LLMs on consumer-grade GPUs
  • Polly is Amazon’s HIPAA- and PCI-compliant, low-latency cloud service that uses NTTS to convert text into lifelike speech
  • Open Assistant (source: LAION-AI/Open-Assistant) is an early stage conversational AI LLM open source project from LAION
  • Prompt Engine is a TypeScript and Python utility library for creating and maintaining prompts for LLMs
  • TensorFlow is Google’s open source library and platform for training and inference of deep neural networks
  • apple/ml-ane-transformers is the reference implementation of the Transformer architecture optimized for Apple Neural Engine (ANE)
  • Semantic Kernel is a lightweight SDK for integrating LLM with conventional programming languages
  • Text generation web UI is a gradio web UI for running Large Language Models like GPT-J 6B, OPT, GALACTICA, LLaMA, and Pygmalion
  • GPTQ is an efficient implementation of the GPTQ algorithm for quantization of GPT models
  • GPTQ-for-LLaMA provides a 4-bit quantization of LLaMA 7B and 13B using GPTQ
  • Visual ChatGPT connects ChatGPT and a series of Visual Foundation Models to enable sending and receiving images during chatting
  • thornewolf/gpt-3-execution provides Python code for an AI assistant that writes and executes code
  • rareranger/chatgpt-linux-assistant is an AI assistant for the Linux command line
  • QAInsights/perfGPT-discord-bot provides Python code for building a GPT chatbot for Discord
  • n3d1117/chatgpt-telegram-bot is a Telegram bot that uses ChatGPT APIs to provide answers
  • joshcho/ChatGPT.el and rksm/org-ai are like GitHub Copilot for Emacs
  • dpayne/CodeGPT.nvim and jackMort/ChatGPT.nvim are like GitHub Copilot for Neovim
  • pyq provides an API to integrate high-level AI-powered features
  • ShareGPT (source: domeccleston/sharegpt) is a Google Chrome extension for quick-sharing of conversations with ChatGPT
  • ChatPDF lets you upload a 200-page PDF and ask questions about anything in the content of the file
  • Flower is an open source federated learning framework
  • OpenMined/PySyft an open source Python library to perform analysis on data hosted in someone else’s server
  • Datapane an open source Python library for building data apps in a Jupyter notebook
  • BlenderGPT is a Blender extension that generates Blender Python code from natural language commands
  • SentenceTransformers is a Python framework for state-of-the-art sentence, text and image embeddings
  • LLMFlows is a framework for building simple, explicit, and transparent LLM (Large Language Model) applications such as chatbots, question-answering systems, and agents
  • guidance programs allow you to interleave generation, prompting, and logical control into a single continuous flow matching how language models processes text
  • LangChain is a framework for composing elements used when interacting with LLMs
  • GPT4All is a LLaMa-based chatbot trained on a massive collection of clean assistant data including code, stories, and dialogue
  • Text2Video-Zero enables zero-shot video generation using a textual prompt combined with guidance from poses or edges and instruction-guided video editing
  • HyDE: Precise Zero-Shot Dense Retrieval without Relevance Labels HyDE zero-shot instructs GPT3 to generate a fictional document and re-encodes it with unsupervised retriever Contriever to search in its embedding space
  • Vocode is an open-source Python library for building voice-based LLM applications
  • Agentflow lets LLMs run workflows specified in JSON files
  • ChainForge is an open-source visual programming environment for prompt engineering, LLM evaluation and experimentation
  • devlooper is a program synthesis agent that autonomously fixes its output by running tests
  • Gdańsk AI is a full stack AI voice chatbot (speech-to-text, LLM, text-to-speech) with integrations to Auth0, OpenAI, Google Cloud API and Stripe - Web App, API and AI

comments powered by Disqus