Directory
Published directory entries matching your search.
Loic Tetrel is a blog about computer science and data science topics. Posts cover subjects such as enabling GPU acceleration in Windows 11 Hyper-V virtual machines, using Android TV without a Google account, and Windows/Linux dual-boot.
Lucent is an adaptation of the Lucid library for PyTorch. It is in a pre-alpha phase and can be installed locally with pip install torch-lucent. Its notebooks can be run via Google Colab or locally with Jupyter.
Implementation of MusicLM, Google's new SOTA model for music generation using attention networks, in Pytorch.
ML/AI Interview Prep is a GitHub-hosted resource collection for machine learning and artificial intelligence technical interview preparation. It includes over 500 interview questions and answers with accompanying code, cheat sheets for NumPy, Pandas, PyTorch, and scikit-learn, and system design content targeted at candidates for roles at companies including Google, Meta, and OpenAI.
A data visualization by Chris Whong showing the activities of one random NYC yellow taxi on a single day in 2013, including where it operated and how much money it made. It uses taxi trip data provided by the Taxi and Limousine Commission via a FOIL request. Trip routes were produced with Google's Directions API and are not assumed accurate.
An archived, read-only GitHub repository holding an open solution to the Google AI Open Images Object Detection Track. It offers starter code, configs and notebooks, with RetinaNet training and references to the Neptune experiment platform; Neptune use is optional and the code can run as a plain Python script. The README notes the competition used Open Images Dataset V4.
Optimum-TPU is a Hugging Face project providing Google TPU optimizations for transformers models, aiming to ease training and inference on TPUs. It supports text generation models including Gemma, Llama2, Llama3 and Mistral, and includes a text-generation-inference backend. The repository was archived in January 2026 and is read-only, with the project in maintenance mode.
Petals is a project for running large language models at home, BitTorrent-style. Users load part of a model and join a network of people serving its other parts. It supports generating text with Llama 3.1, Mixtral, Falcon and BLOOM and fine-tuning them on a consumer-grade GPU or Google Colab.
pixray is an image generation system combining perception engines, CLIP-guided GAN imagery and CLIPDraw. It is a Python library and command line utility that can also run in Google Colab notebooks or inside Docker using Cog. Installation requires a recursive git clone and a fresh Python 3.8 environment.
Plus is an AI assistant for PowerPoint and Google Slides that helps users create, refine and present presentations. It can generate slides from a prompt or an uploaded file, rewrite and reformat content, add images and charts, and work with custom templates and preset libraries.
PricePerToken is a free comparison site for LLM API pricing. It presents per-token costs for models from providers including OpenAI, Anthropic, Google and Mistral, alongside model listings with context windows and benchmark figures. Pages cover individual models such as Amazon Nova Lite 1.0 and Nova Micro 1.0, whose pricing and context windows are stated.
This is the official project page for the ICML 2024 research paper 'Improving Factuality and Reasoning in Language Models through Multiagent Debate' by MIT and Google Brain researchers. The paper introduces a multi-round debate approach where multiple language model instances propose answers, critique each other, and debate to improve reasoning and factual accuracy while reducing hallucinations.
pytessarct is a Python wrapper for the Google Tesseract optical character recognition tool, hosted as a public repository on GitHub. It enables developers to incorporate Tesseract's text recognition capabilities into Python projects.
Q, ChatGPT for Slack is an officially approved Slack app powered by OpenAI. Users interact with it using @Q in channels or without @Q in direct messages, and it offers plugins such as Google search and document search. It includes a 14-day free trial with no credit card required.
QuickCreator describes itself as an AI content team for solo founders that produces articles in the user's voice intended to rank on Google and be cited by ChatGPT, Perplexity, and Gemini. Plans range from $29 per month to custom enterprise pricing, with a 7-day free trial.
Read AI is an AI meeting notetaker that works across Microsoft Teams, Google Meet, and Zoom. It generates meeting summaries, transcripts, and action items, and offers search across meetings, emails, and messages. It is also available as a Claude Connector, ChatGPT App, and MCP Server.
Refinder AI boosts productivity in Slack and Google Chat by automating tasks, managing apps, and providing instant access to knowledge.
Scholarcy is a tool for summarizing and organizing research, built around AI-powered tools described as made specifically for academic papers. It converts long texts into interactive summary flashcards highlighting key information, and accepts imports including PDFs, book chapters, articles, Zotero, Google Drive and YouTube. Summaries can be enhanced and exported to other apps and citation managers.
Scribbl is an AI meeting note taker for Google Meet, Zoom and Microsoft Teams in which no bot joins the call. After each meeting it provides AI notes, action items, a transcript with speaker names and the video. The free plan includes 10 meetings a month with no limit on meeting length.
SlidesAI is a web-based tool that turns text, notes, documents, PDFs and prompts into presentations. It works within Google Slides, PowerPoint, ChatGPT and a browser-based web app. The Basic plan costs $0/month and includes 12 presentations per year. The Pro plan costs $10.00/month billed at $120/year and adds document upload.
SlidesWizard is an AI tool that creates presentations and pitch decks from a topic or a pasted outline, adding images and speaker notes. It works with PowerPoint and Google Slides. A free option is offered, and a pay-per-use option costs $2.50 per presentation.
A public GitHub repository described as a repo for jailbreaking various LLMs, mainly Claude. It includes a Jailbreak-Guide folder and a README stating its mission to test the boundaries of ChatGPT, Claude, Gemini, Grok and other LLMs. The README credits contributors and says the repo was compiled via Google Jules.
Versão em português do notebook para rodar a Web UI do Stable Diffusion no Google Colab de graça.
TensorFlow Fold is a library for creating TensorFlow models that consume structured data, where the computation graph depends on the structure of the input data. It implements dynamic batching, transforming batches of arbitrarily shaped computation graphs into a static computation graph. TensorFlow Fold is not an official Google product.
