franz heffa
Date d'abonnement : 2025
Date d'abonnement : 2025
Train more powerful models with a single GPU. In this course, you will learn how hardware can speed up model training and the key considerations when training models on a GPU. First, you will learn how to estimate the number of computations and the amount of computer memory required to train large neural networks. You will then discover techniques for reducing the computing and memory requirements when training a model. Techniques which you will apply for fine-tuning a Gemma model with 4 billion parameters. Finally, you will consider the potential environmental impacts of machine learning, with a focus on where questions of energy, water, and e-waste intersect with justice and equity.
Mettez en pratique vos connaissances sur les agents en créant, configurant et exécutant votre premier agent IA à l'aide d'Agent Development Kit (ADK) de Google. Dans ce cours pratique, vous configurerez un environnement de développement ADK complet, créerez des agents avec du code Python et une configuration YAML, et les exécuterez sur plusieurs interfaces. Vous découvrirez également les paramètres essentiels qui définissent le comportement de l'agent, en appliquant dans du code fonctionnel ce que vous avez appris dans le cours 1.
Découvrez comment utiliser Agent Development Kit (ADK). Ce cours présente le framework Open Source d'ADK et montre comment passer du prompt engineering simple à une approche de développement logiciel structurée, axée sur le code, adaptée aux systèmes multi-agents de niveau entreprise.
Unleash the power of language models with fine-tuning. In this course, you will learn how to adjust a pre-trained model to a specific task. You will start with full-parameter fine-tuning using a small language model. To tune larger models like Gemma, you will learn parameter-efficient techniques with a focus on LoRA. Finally, you will be briefly introduced to reinforcement learning as an alternative to supervised fine-tuning (SFT). You will also explore how AI is imagined and made sense of in cultural contexts. You will consider why responsible AI is not just about technical safety but also about building governance systems that reflect community values and protect the public interest.
Les agents IA représentent une évolution majeure par rapport aux grands modèles de langage traditionnels (LLM) : au lieu de simplement générer des solutions textuelles, ils peuvent également agir de manière autonome pour les exécuter. Ce cours présente les principes de base des agents IA, leurs différences avec les API LLM et les cas où ils apportent une réelle valeur. Basé sur le livre blanc de Google consacré aux agents, il fournit les bases théoriques nécessaires avant d'écrire vos premières lignes de code d'agent. Il est idéal pour les développeurs, les architectes et les décideurs techniques qui souhaitent comprendre les systèmes d'IA à travers le prisme du comportement autonome et orienté vers des objectifs, plutôt que simplement comme des générateurs de texte. Rejoindre le forum de la communauté pour poser des questions et discuter
Complete the advanced Google DeepMind: Train A Small Language Model skill badge by completing this course to demonstrate skills in the following: formulating real-world language model research problems; building a simple tokenizer; preparing a dataset for training a transformer language model; running the training loop of a small language model. Access this lab at no-cost by signing up for the no-cost subscription. Receive 35 free credits each month!
Gemini pour Google Workspace est un module complémentaire qui fournit aux clients des fonctionnalités d'IA générative dans Google Workspace. Dans ce parcours de formation, vous allez découvrir les principales fonctionnalités de Gemini et comment elles peuvent servir à améliorer la productivité et l'efficacité dans Google Workspace.
Ce cours présente le concept d'IA responsable et les principes associés. Il met en avant des techniques permettant d'identifier des données équitables ou biaisées, et de limiter les biais lors de l'utilisation de l'IA/du ML. Vous découvrirez des méthodes pratiques et des outils pour mettre en place de bonnes pratiques d'IA responsable à l'aide des produits Google Cloud et des outils Open Source.
This course is for developers interested in learning how to use TPUs for inference—from architecture to deployment, and how to solve common implementation challenges.
This course is designed for developers looking to build an optimized AI inference stack on Google Cloud. Whether you’re working with GPUs or TPUs, you’ll explore the fundamental components of an inference stack, learn design principles for maximizing performance and reliability, and explore practical techniques to take your workloads from 0 to 1.
Ce cours apporte aux professionnels du machine learning les techniques, les bonnes pratiques et les outils essentiels pour évaluer les modèles d'IA prédictive et générative. L'évaluation des modèles est primordiale pour s'assurer que les systèmes de ML fournissent des résultats fiables, précis et de haut niveau en production. Les participants acquerront une connaissance approfondie de diverses métriques et méthodologies d'évaluation, ainsi que de leur application appropriée dans différents types de modèles et tâches. Le cours mettra l'accent sur les défis uniques posés par les modèles d'IA générative et proposera des stratégies pour les relever efficacement. Grâce à la plate-forme Vertex AI de Google Cloud, les participants apprendront à implémenter des processus d'évaluation rigoureux pour la sélection, l'optimisation et la surveillance continue des modèles.
Dans ce cours, vous allez acquérir les connaissances et les outils nécessaires pour identifier les problématiques uniques auxquelles les équipes MLOps sont confrontées lors du déploiement et de la gestion de modèles d'IA générative. Vous verrez également en quoi Vertex AI permet aux équipes d'IA de simplifier les processus MLOps et de faire aboutir leurs projets d'IA générative.
In this video, you'll learn to use Gemini in Google Sheets for advanced data analysis. You'll learn how to create data visualizations, like a scatter plot, using a simple prompt. You'll also learn how to ask Gemini to analyze your data for strategic insights and recommendations, such as how to save money.
In this video, you'll learn to use Gemini in Google Sheets to automatically generate organized trackers from unstructured data. You'll learn how to write a prompt that references other Google Drive files using the "@" symbol. You'll also see how Gemini builds a fully formatted, color-coded table with drop-down menus based on the data in your notes, saving you significant manual effort.
In this video, you'll learn to use Gemini in Google Workspace to manage your calendar from your Gmail inbox. You'll learn how to ask Gemini about your schedule in the side panel and how to create new events without switching apps. You'll also learn how to use Gemini's one-click scheduling feature when it automatically detects event details in an email.
In this video, you'll learn to use Gemini in Google Slides to build polished presentations efficiently. You'll see how to create slides by referencing content directly from your Google Drive files using the "@" symbol. You'll also learn to generate custom, professional images from a simple text prompt to visually enhance your presentation.
In this video, you'll learn to use the "Ask Gemini" feature in Google Sheets to manage your data with natural language. You'll see how to ask Gemini to create dropdowns, highlight sales data, create a pivot table, filter by month, and sort by revenue.
In this video, you'll learn to use the =AI() function in Google Sheets to automate your work. You'll see how to generate text like slogans, summarize paragraphs of customer feedback, and categorize data by classifying inquiries and analyzing sentiment.
In this video, you'll learn to create "what if" scenarios using AI's reasoning. You'll see how to set a "before" scene with a person holding a cake, prompt an action by asking "what would happen if they tripped?", and let the AI generate the plausible "after" image of the cake falling.
In this video, you'll learn the "creative mashup artist" technique for combining images. You'll see how to generate a subject and a scene as two separate images, and then use a final blending prompt to fuse the astronaut from the first image and the court from the second into one new picture.
In this video, you'll learn to art direct your images in the Gemini app. You'll see how to use multi-turn editing to furnish an empty room step-by-step, and how to "remix" a photo by applying the color and texture of a butterfly's wings to a pair of rainboots.
In this video, you'll learn to create imaginative portraits in the Gemini app that maintain your likeness. You'll see how to use the "Subject, Action, Scene, Style" formula for better prompts, transport yourself into a vintage photo, and combine a photo of yourself and your dog into one new image.
In this video, you'll learn to use NotebookLM to manage a large-scale research project. You'll see how to upload all your documents to get summaries, ask deeper strategic questions to find key insights, and use the audio feature to listen to your findings on the go.
In this video, you'll learn to use NotebookLM to analyze raw data and find compelling stories. You'll see how to ground the AI in your specific documents, prompt it to generate newsworthy PR claims, and ask it to find hidden correlations between different data points to uncover new insights.
In this video, you'll learn to use an AI assistant to streamline large-scale event planning. You'll see how to prompt an AI to brainstorm off-site ideas, create fair groups for icebreakers, build a shuttle schedule based on arrival times, and analyze survey feedback for gains and losses.
In this video, you'll learn to use Gemini Deep Research to make complex purchasing decisions with confidence. You'll see how to write a nuanced, personal prompt about your specific needs and receive a synthesized, comparative report that cuts through the clutter of online reviews.
In this video, you'll learn to create an animated bar chart race using Gemini Canvas. You'll see how to write a prompt to visualize data over time and then use a follow-up prompt to refine the animation, such as slowing it down for a presentation.
In this video, you'll learn to create complex animated art using Gemini Canvas. You'll see how to write a descriptive prompt to generate a kinetic typography animation with distortion effects, and then use a follow-up prompt to completely change its color style.
In this video, you'll learn to build a reusable, custom "Study Guide Gem." You'll see how to write the core instructions for your Gem once, and then use it again and again to transform messy notes and articles into a perfectly structured study guide with a summary, outline, and glossary.
In this video, you'll learn to turn your study guides into interactive quizzes using Gemini Canvas. You'll see how to upload your notes and use a simple prompt to generate a quiz with multiple choice and fill-in-the-blank questions that provide immediate feedback.
In this video, you'll learn to use Gemini Deep Research to conduct competitive marketing analysis. You'll see how to submit a research prompt, approve the personalized research plan Gemini creates, and receive a detailed, synthesized report on competitor campaigns to inform your strategy.
In this video, you'll learn to build a working prototype of a personalized weather app using Gemini Canvas. You'll see how to specify the cities for a 7-day forecast and then use a follow-up prompt to change the app's theme to a minimal, dark mode interface.
In this video, you'll learn to create an interactive 3D simulation using Gemini Canvas. You'll see how to build a 3D model of the solar system from a single sentence and then use follow-up prompts to add interactive and educational features, like clickable planets and fun facts.
In this video, you'll learn to build a custom personal utility app using Gemini Canvas. You'll see how to write a detailed prompt to create a Pomodoro timer tailored to your specific needs, and then use follow-up prompts to adjust its visual design.
In this video, you'll learn to use the Guided Learning experience in Gemini to build a deep understanding of any topic. You'll see how to start a conversation, engage with Gemini's probing questions, and use interactive tools like diagrams and quizzes to guide your learning journey.
In this video, you'll learn to use a simple text prompt in Gemini Canvas to build a playable game. You'll see how to create a Tic-Tac-Toe game, use a follow-up prompt to customize its theme to a retro 8-bit style, and share the final game with a link.
In this video, you'll learn to build a working music synthesizer using a simple text prompt in Gemini Canvas. You'll see how to specify controls like waveform, attack, and sustain, and then interact with the generated instrument to create and experiment with your own unique sounds.
In this video, you'll learn to use creative prompts in Gemini to make complex research fun and easy to understand. You'll see how to ask Gemini to research a topic and present the findings in a creative format, like a sports scouting report and a head-to-head bracket, to better visualize and compare your options.
This video covers how to use the Video Overviews feature in NotebookLM to automatically generate a short explainer video based on your source documents.
This video covers how to use the 'Discover Sources' feature in NotebookLM to find and import relevant web-based sources directly into your research project.
This video covers how to use the Mind Maps feature in NotebookLM to automatically create a visual representation of your sources, helping you understand connections and key concepts.
In this video, you'll learn to use the Create menu in Gemini Canvas to instantly convert a block of text into a variety of visual formats. You’ll see how to take a brainstormed marketing campaign and turn it into both a shareable infographic and an internal webpage with just one click.
In this video, you'll learn to use Gemini Canvas to turn a simple drawing into a working app. You'll upload a sketch, use a simple prompt to generate code, watch a live preview build itself, and make iterative changes to your app prototype using natural language.
This video covers how NotebookLM's Reports feature dynamically suggests formats to help you create customized, trustworthy analyses of your documents with ease.
This video covers how to build a personalized "Work with Me" agent using Gemini Gems, which helps streamline foundational feedback and makes your meetings more strategic and efficient.
AI Boost Bites is a video series designed to help you leverage Google's AI tools in your daily work. Each episode, under 10 minutes, features a quick video demonstrating a real-world AI use case or topic. After the video, you'll get a challenge to apply what you've learned. It's an easy, interactive way to boost your AI skills and improve your productivity.
This video covers how to use Gemini in Gmail to draft new emails, refine their tone, respond with context from Drive files, and use smart reply suggestions.
This video covers how to use the 'Help me create' feature in Google Docs to generate a complete, formatted document by referencing content from other files in your Drive.
This video covers how to use Gemini in Gmail to summarize emails, find information, and draft replies, helping you manage your inbox more efficiently.
This video covers how to use Gemini in Slides to automatically generate meeting recaps and draft follow-up emails, which can streamline your post-meeting workflow and save you time.
This video covers how you can leverage Gemini's advanced AI capabilities within Google Sheets to effortlessly pull data and generate insights in minutes, all without the need for any technical or coding background.
This video will cover how to leverage Gemini Gems to create authentic social media posts in your leader's unique voice. Learn to overcome the challenge of scaling executive social presence by training a Gem with writing samples and clear instructions. Discover how to generate engaging posts quickly, saving time while amplifying thought leadership and ensuring authenticity.
This video covers how you can create your own Brevity Gem to summarize and transform messy notes or long documents into clear, concise, executive-ready summaries.
This video covers how to use Gemini and Apps Script to automate manual tasks across Google Workspace. You'll learn to prompt Gemini to generate Apps Script code that automatically drafts email reminders in Google Sheets for tasks not marked 'Complete.' Automate your workflow with little to no technical expertise, freeing up time for more important work and eliminating manual follow-ups.
This video covers how you can use Gemini to summarize long documents in Google Workspace, so you can quickly get the information you need and save time. You'll learn how to use Gemini to summarize entire documents or just selected text, as well as how to use Gemini in Drive to summarize across multiple files.
This video covers how you can leverage Notebook LM to "eat the frog" on your to-do list by automating complex tasks like summarizing legislation and mapping services, saving you hours of work.
This video covers how to eliminate tedious manual data entry using Gemini. Learn how to take a picture or screenshot of data (from PDFs, paper, or images) and prompt Gemini to instantly convert it into a structured Google Sheet. Discover this simple hack to save countless hours transcribing data, turning Gemini into your personal data entry assistant. Just snap, prompt, and export!
AI Boost Bites is a video series designed to help you leverage Google's AI tools in your daily work. Each episode, under 10 minutes, features a quick video demonstrating a real-world AI use case or topic. After the video, you'll get a challenge to apply what you've learned. It's an easy, interactive way to boost your AI skills and improve your productivity.
This video will cover how to use NotebookLM to gather and analyze publicly available information, combine it with internal documents, and extract key competitive insights.
This video covers how to personalize your Gemini results in Google Workspace. Learn to incorporate documents and research papers directly into your prompts using the "@" symbol to get more targeted and relevant AI output tailored to your needs.
This video will cover how you can leverage Gemini's advanced AI capabilities in Google Docs to brainstorm ideas, draft various marketing content, and collaborate with your team.
In this video, you'll learn pro tips for conducting thorough competitive research with Gemini Deep Research. You'll see how to enhance a basic prompt by assigning a persona, specifying a table format for the output, and directing the AI to search specific sources like Reddit for richer insights.
This video covers how to use NotebookLM for common marketing tasks like analyzing customer feedback, conducting market research, and generating content ideas.
This video covers how to use NotebookLM as a personal research assistant by adding sources, asking questions, and generating new content formats based on your documents.
This video covers how to create a 'project notebook' in NotebookLM by adding all relevant sources to build a central, searchable knowledge hub for your team.
This video covers five key ways to use Google's AI tools, including Gemini in Workspace, the Gemini app, and NotebookLM, to enhance your daily productivity.
This video covers prompt engineering fundamentals for effective AI communication. Learn a simple framework (Persona, Task, Context, Format) to craft clear prompts, getting better, faster results from Gemini in Google Workspace. Discover how to use natural language, be specific, and iterate for optimal AI assistance.
Vous souhaitez utiliser AI Hypercomputer ? Ce cours est conçu pour vous aider à vous lancer. Nous aborderons les principes de base de cet outil et comment il aide l'IA dans ses charges de travail. Vous découvrirez les différents composants d'un hypercalculateur, tels que les GPU, les TPU et les CPU, et apprendrez à choisir la méthode de déploiement la mieux adaptée à vos besoins.
"Concepts fondamentaux de Google Cloud : infrastructure de base" présente les concepts et les termes à connaître pour utiliser Google Cloud. À travers des vidéos et des ateliers pratiques, il décrit et compare la plupart des services Google Cloud de calcul et de stockage, ainsi que des outils importants de gestion des ressources et des règles.
Kubernetes est le système d'orchestration de conteneurs le plus populaire, et Google Kubernetes Engine a été conçu spécifiquement pour les déploiements gérés de Kubernetes dans Google Cloud. Dans ce cours de niveau avancé, vous allez suivre des ateliers pratiques pour apprendre à configurer les images et les conteneurs Docker, ainsi qu'à déployer des applications Kubernetes Engine opérationnelles. Vous allez également acquérir les compétences nécessaires pour intégrer l'orchestration de conteneurs à votre propre workflow. Vous cherchez un atelier challenge pratique pour démontrer vos compétences et valider vos connaissances ? Suivez cet atelier challenge complémentaire après avoir terminé ce cours et le cours Déployer des applications Kubernetes sur Google Cloud pour recevoir un badge numérique Google Cloud exclusif.
Ce cours aide les participants à créer un plan de formation pour l'examen de certification qui permet de devenir Professional Cloud Architect. Ils découvriront l'ampleur et le champ d'application des domaines abordés lors de l'examen. Ils détermineront s'ils sont prêts à passer l'examen et créeront leur propre plan de formation.
In this Google DeepMind course you will focus on the training process for machine learning models. You will learn how to spot and mitigate issues when training a model, such as overfitting and underfitting. In practical coding labs, you will implement and evaluate the multilayer perceptron for simple classification tasks. This will provide insights into the mechanics of training a neural network model and the backpropagation algorithm. Research case studies will demonstrate how neural networks power real-world models. Additionally, you will consider the broader social impacts of innovation by looking beyond immediate benefits to anticipate potential risks, safety concerns, and further-reaching societal consequences.
In this Google DeepMind course you will learn how to prepare text data for language models to process. You will investigate the tools and techniques used to prepare, structure, and represent text data for language models, with a focus on tokenization and embeddings. You will be encouraged to think critically about the decisions behind data preparation, and what biases within the data may be introduced into models. You will analyze trade-offs, learn how to work with vectors and matrices, how meaning is represented in language models. Finally, you will practice designing a dataset ethically using the Data Cards process, ensuring transparency, accountability, and respect for community values in AI development.
In this Google DeepMind course, you will learn the fundamentals of language models and gain a high-level understanding of the machine learning development pipeline. You will consider the strengths and limitations of traditional n-gram models and advanced transformer models. Practical coding labs will enable you to develop insights into how machine learning models work and how they can be used to generate text and identify patterns in language. Through real-world case studies, you will build an understanding around how research engineers operate. Drawing on these insights you will identify problems that you wish to tackle in your own community and consider how to leverage the power of machine learning responsibly to address these problems within a global and local context.
In this Google DeepMind course you will discover the mechanisms of the transformer architecture. You will investigate how transformer language models process prompts to make context-sensitive next-token predictions. Through practical activities you will explore the attention mechanism, visualize attention weights, and encounter advanced concepts like masked attention and multi-head attention. You will also learn other techniques that are necessary to build neural networks that are well-suited to be used as language models. Finally, through activities on values, stakeholder mapping and community engagement, you will practice concrete tools for ensuring AI projects are developed with communities, not just for them.
Ce cours a été conçu pour préparer les administrateurs Google Workspace à résoudre les problèmes courants liés à Google Workspace. Les participants s'entraîneront à diagnostiquer et résoudre des problèmes dans Gmail, Agenda et Drive, et à utiliser la console d'administration. Ils apprendront également à analyser des journaux d'audit pour résoudre des problèmes de sécurité, ainsi qu'à recueillir des informations et à exploiter les ressources disponibles pour résoudre et signaler des problèmes techniques.
Ce cours apprend aux participants à sécuriser leur environnement Google Workspace. Les participants mettront en place des règles de mot de passe sécurisées ainsi que la validation en deux étapes pour gérer l'accès des utilisateurs. Ils utiliseront ensuite l'outil d'investigation de sécurité afin d'identifier et de résoudre les problèmes de sécurité de manière proactive. Ensuite, ils géreront l'accès aux applications tierces et les appareils mobiles afin d'assurer la sécurité. Pour finir, les participants appliqueront les mesures de sécurité des e-mails et de conformité pour protéger les données organisationnelles.
Ce cours apporte aux participants des compétences de gouvernance des données dans leur environnement Google Workspace. Les participants étudieront l'utilisation de règles de protection contre la perte de données dans Gmail et Drive afin de prévenir les fuites de données. Ils apprendront ensuite à utiliser Google Vault pour la conservation et la récupération des données. Ils découvriront ensuite comment configurer les régions de données et les paramètres d'exportation afin de se conformer à la réglementation. Enfin, ils verront comment classifier les données à l'aide d'étiquettes pour améliorer l'organisation et renforcer la sécurité.
Ce cours a été conçu pour présenter en détail les services principaux de Google Workspace. Les participants y apprendront à activer, désactiver et configurer les paramètres de ces services, dont Gmail, Agenda, Drive, Meet, Chat et Docs. Ensuite, ils découvriront comment déployer et gérer Gemini dans l'intérêt de leurs utilisateurs. Enfin, les participants examineront des cas d'utilisation d'AppSheet et d'Apps Script pour apprendre à automatiser des tâches et étendre les fonctionnalités des applications Google Workspace.
Ce cours a été conçu pour présenter la gestion des ressources et des utilisateurs dans Google Workspace. Les participants y apprendront à configurer des unités organisationnelles pour répondre aux besoins de leur organisation, et à gérer différents types de groupes Google. Ils développeront également une expertise dans la gestion des paramètres de domaine dans Google Workspace. Pour finir, ils apprendront à maîtriser l'optimisation et la structuration des ressources dans leur environnement Google Workspace.
Le cours "IA générative : au-delà du chatbot" est le premier du parcours de formation "Leader en IA générative" et n'a aucun prérequis. Ce cours vise à approfondir votre compréhension de base des chatbots afin de révéler le véritable potentiel de l'IA générative pour votre entreprise. Vous découvrirez des concepts tels que les modèles de fondation et le prompt engineering (ingénierie des requêtes), qui sont essentiels pour exploiter toute la puissance de l'IA générative. Ce cours vous aidera également à identifier les facteurs à prendre en compte pour développer une stratégie d'IA générative efficace pour votre entreprise.
This video covers how NotebookLM can revolutionize customer insight gathering from call or chat transcripts. You'll learn to upload PDF transcripts of hundreds of conversations (even multilingual ones!) and quickly extract key themes, trending topics, and actionable insights without listening for hours. Discover how to save findings, share notebooks, and even generate interactive podcast summaries of your data.
This video covers how to create your own Gemini Gems, advanced AI capabilities that can automate repetitive tasks and supercharge your productivity.
Ce cours présente une solution de génération augmentée par récupération (RAG) dans BigQuery permettant de réduire les hallucinations de l'IA. Il décrit un workflow RAG qui couvre la création d'embeddings, la recherche dans un espace vectoriel et la génération de réponses améliorées. Il explique aussi les raisons conceptuelles derrière ces étapes et leur implémentation pratique avec BigQuery. À la fin du cours, les participants seront à même de créer un pipeline de RAG à l'aide de BigQuery et de modèles d'IA générative tels que Gemini, ainsi que des modèles d'embeddings pour traiter leurs propres cas d'hallucinations de l'IA.