Sam Altman a pris la parole lors d'une conférence sur les progrès futurs de l'IA à l'Université de Stanford. L'occasion de parler du développement de GPT-5 et le patron d'OpenAI le promet : la prochaine évolution de ChatGPT sera grande.
[Lire la suite]Envie de rejoindre une communauté de passionnés ? Notre Discord vous accueille, c’est un lieu d’entraide et de passion autour de la tech.
Google est encore très prudent concernant l'utilisation de l'intelligence artificielle générative dans ses produits. Mais cette prudence pourrait-elle finalement causer sa chute ? OpenAI, l'éditeur de ChatGPT, pourrait bientôt lancer un moteur de recherche basé sur sa technologie ChatGPT, qui pourrait concurrencer directement Google.
[Lire la suite]Utilisez-vous Google News (Actualités en France) ? Vous pouvez suivre vos médias favoris. Suivez Frandroid sur Google News (et Numerama).
Apparu mystérieusement sur un site de comparaison des grands modèles de langage, le modèle gpt2-chatbot intrigue la communauté de l'intelligence artificielle (IA). Supposément capable de résoudre des problèmes inabordables pour GPT-4, il pourrait être un prototype d'un futur modèle OpenAI. Sam Altman, le patron de l'entreprise, ne cache pas son amusement.
Vous avez entendu parler de GPT2-chatbot ?
C’est un modèle de langage un peu mystérieux, accessible uniquement sur le site https://chat.lmsys.org, qui semble avoir des super pouvoirs dignes de ChatGPT. Mais attention, suspense… Personne ne sait d’où il sort ! Ce chatbot anonyme fait tourner les têtes cette semaine après être devenu disponible sur un important site de référence pour les grands modèles de langage, LMSYS Org. Beaucoup considèrent qu’il a à peu près les mêmes capacités que GPT-4 d’OpenAI, ce qui le place dans une rare catégorie de modèles d’IA que seule une poignée de développeurs dans le monde a pu atteindre.
Quand on lui pose la question, ce petit malin de GPT2-chatbot clame haut et fort qu’il est basé sur l’archi de GPT-4 sauf que voilà, ça colle pas vraiment avec son blaze GPT-2…
Les communautés d’IA en ligne se sont emballées au sujet de l’anonyme gpt2-chatbot. Un utilisateur de X affirme que gpt2-chatbot a presque codé un clone parfait du jeu mobile Flappy Bird. Un autre utilisateur de X dit qu’il a résolu un problème de l’Olympiade internationale de mathématiques en un seul coup. Sur de longs fils Reddit, les utilisateurs spéculent sauvagement sur les origines de gpt2-chatbot et se disputent pour savoir s’il provient d’OpenAI, de Google ou d’Anthropic. Il n’y a aucune preuve de ces affirmations, mais les tweets de Sam Altman, PDG d’OpenAI, et d’autres cadres n’ont fait que jeter de l’huile sur le feu.
Mise à jour : De nouvelles informations importantes sont apparues concernant GPT2-chatbot :
Tout cela va clairement dans le sens de l’hypothèse selon laquelle GPT2-chatbot serait bien un nouveau modèle GPT d’OpenAI, probablement une version préliminaire de GPT-4.5. Les performances sont en effet un cran au-dessus de GPT-4 tout en restant dans la même lignée.
L’accès à GPT2-chatbot est actuellement limité à 8 messages par jour et par utilisateur en mode « tchatche directe ». Pour continuer après, il faut passer en mode « Battle ». Les restrictions plus importantes que pour GPT-4 suggèrent que le modèle a un coût de calcul plus élevé.
Malheureusement, suite à un trafic trop important, LMSYS a dû temporairement désactiver l’accès à GPT2-chatbot. Affaire à suivre donc pour découvrir l’identité réelle de ce mystérieux modèle et les plans d’OpenAI à son sujet. Une version plus large sera-t-elle bientôt diffusée ? Réponse dans les prochaines semaines !
Prêt à tester les talents cachés de GPT2-chatbot ?
Si un jour, ça remarche, direction https://chat.lmsys.org, sélectionnez « gpt2-chatbot », cliquez sur « Chat » et c’est parti mon kiki !
Vous aurez le droit à 8 messages gratos en mode « tchatche directe » et après, faut passer en mode « Battle » pour continuer à jouer. Un petit conseil : pensez à repartir d’une page blanche en cliquant sur « New Round » à chaque fois que vous changez de sujet, sinon il risque de perdre le fil.
On verra bien dans quelques semaines quelle théorie sortira gagnante de ces discussions. Il y a très peu d’informations disponibles sur gpt2-chatbot pour l’instant mais il semble clair qu’un acteur majeur est derrière ce modèle IA.
De nouveaux ennuis arrivent pour OpenAI, la société derrière ChatGPT. L'activiste autrichien Max Schrems, très actif contre les géants du net dès qu'il est question de données personnelles, a lancé une procédure contre le créateur du célèbre chatbot.
Dubaï a été récemment frappé par un épisode météorologique d'une ampleur rare, avec des précipitations exceptionnelles -- au point de provoquer des inondations. Une hypothèse a suggéré que le violent orage qui a frappé l'émirat était dû à un ensemencement de nuages qui a mal tourné. C'est toutefois improbable.
Un ancien animateur des célèbres studios Pixar estime que, pour l'instant, les outils capables de générer des vidéos grâce à l'IA, ne seraient pas encore assez pertinents. La nécessité de retravailler les plans empêche ces programmes de percer dans le monde du cinéma.
« La France est le premier pays européen à inaugurer une IA 100 % souveraine et à la mettre au service de nos services publics », a fait savoir le Premier ministre Gabriel Attal le 23 avril. Cette IA, c'est Albert, un chatbot pour appuyer l'administration.
Le drame épique, situé dans un Japon féodal implacable, vient de tirer sa révérence avec un ultime épisode sur Disney+. Pour prolonger le plaisir, voici donc 6 séries similaires à Shōgun, à voir en streaming : Succession, Blue Eye Samurai, Vikings, Le Temps des Samouraïs, Tokyo Vice ainsi que Giri/Haji.
This blog is co-authored by Gordan Markuš, Canonical and Kumar Sankaran, Ventana Micro Systems
RISC-V, an open standard instruction set architecture (ISA), is rapidly shaping the future of high-performance computing, edge computing, and artificial intelligence. The RISC-V customizable and scalable ISA enables a new era of processor innovation and efficiency. Furthermore, RISC-V democratizes innovation by allowing new companies to develop their own products on its open ISA, breaking down barriers to entry and fostering a diverse ecosystem of technological advancement.
By fostering a more open and innovative approach to product design, the RISC-V technology vendors are not just a participant in the future of technology; they are a driving force behind the evolution of computing across multiple domains. Its impact extends from the cloud to the edge:
One of the unique products that can be designed with RISC-V ISA are chiplets. Chiplets are smaller, modular blocks of silicon that can be integrated to form a larger, more complex chip. Instead of designing a single monolithic chip, a process that is increasingly challenging and expensive at cutting-edge process nodes, manufacturers can create chiplets that specialize in different functions and combine them as needed. RISC-V and chiplet technology is empowering a new era of chip design, enabling more companies to participate in innovation and tailor their products to specific market needs with unprecedented flexibility and cost efficiency.
Canonical makes open source secure, reliable and easy to use, providing support for Ubuntu and a growing portfolio of enterprise-grade open source technologies. One of the key missions of Canonical is to improve the open source experience across ISA architectures. At the end of 2023, Canonical announced joining the RISC-V Software Ecosystem (RISE) community to support the open source community and ecosystem partners in bringing the best of Ubuntu and open source to RISC-V platforms.
As a part of our collaboration with the ecosystem, Canonical has been working closely with Ventana Micro Systems (Ventana). Ventana is delivering a family of high-performance RISC-V data center-class CPUs delivered in the form of multi-core chiplets or core IP for high-performance applications in the cloud, enterprise data center, hyperscale, 5G, edge compute, AI/ML and automotive markets.
The relationship between Canonical and Ventana started with a collaboration on improving the upstream software availability of RISC-V in projects such as u-boot, EDKII and the Linux kernel.
Over time, the teams have started enabling Ubuntu on Ventana’s Veyron product family. Through the continuous efforts of this partnership Ubuntu is available on the Ventana Veyron product family and as a part of Ventana’s Veyron Software Development Kit (SDK).
Furthermore, the collaboration extends to building full solutions for the datacenter, HPC, AI/ML and Automotive, integrating Domain Specific Accelerators (DSAs) and SDKs, promising to unlock new levels of performance and efficiency for developers and enterprises alike. Some of the targeted software stacks can be seen in the figure below.
Today, Ventana and Canonical collaborate on a myriad of topics. Together through their joint efforts across open source communities and as a part of RISC-V Software Ecosystem (RISE), Ventana and Canonical are actively contributing to the growth of the RISC-V ecosystem. We are proud of the innovation and technology leadership our partnership brings to the ecosystem.
Ubuntu is the reference OS for innovators and developers, but also the vehicle to enable enterprises to take products to market faster. Ubuntu enables teams to focus on their core applications without worrying about the stability of the underlying frameworks. Ventana and the RISC-V ecosystem recognise the value of Ubuntu and are using it as a base platform for their innovation.
Furthermore, the availability of Ubuntu on RISC-V platforms not only allows developers to prototype their solutions easily but provides a path to market with enterprise-grade, secure and supported open source solutions.Whether it’s for networking offloads in the data center, training AI models in the cloud, or running AI inference at the edge, Ubuntu is an established platform of choice.
Contact Canonical to bring Ubuntu and open source software to your RISC-V platform.
We are excited to share that Canonical participates in America Digital Congress in Santiago, Chile, for the first time ever. It’s one of the leading events in the region about digital transformation bringing together VPs and experts from the most relevant global tech companies.
Canonical, the publisher of Ubuntu, provides open source security, support and services. In addition to the OS, Canonical offers an integrated data and AI stack. With customers that include top tech brands, emerging startups, governments and home users, Canonical delivers trusted open source for everyone.
Join us at the booth A31 to learn how Canonical can support your digital transformation journey securely and cost-efficiently.
Juan Pablo Noreña, Canonical Cloud Field Software Engineer, is delighted to be speaking at America Digital Congress about digital transformation and AI. In this talk, he will explore the significant benefits of introducing open source solutions in all stages of the infrastructure implementation process, from virtualization to AI platforms.
Juan Pablo will also showcase how this approximation improves security, reduces costs in the infrastructure life cycle, and makes them predictable, offering companies a competitive advantage in the market.
Key topics:
At Canonical, we provide the services our partners need to ensure their hardware and software works optimally with the Ubuntu platform. We operate a range of partner programmes, from essential product certification to strategic collaboration, help with QA and long-term strategic alliances. For technology customers, this has created a thriving market of suppliers with Ubuntu expertise.
Are you interested to learn more about our partner programmes? Talk to the team at the booth or visit our partner webpage.
Come visit us at the booth to learn how Canonical could support you in the digital transformation journey. Check out our Data and AI offerings to learn more about our solutions.
Today, we bring you a major update to the Nextcloud AI Assistant with Context Chat, Context Write, GPU acceleration, the ability to split off the Nextcloud server & AI server, and MUCH more! And for those without $100K GPU’s (you know who you are), we have also incredible news: we are working with several prominent European hosting companies, including IONOS, OVHcloud and plusserver to deliver AI-as-a-Service solutions that respect your privacy and digital sovereignty! For our US users and customers, there are already several companies offering AI-as-a-Service and some we support (OpenAI in particular) while others are in the works.
With a nearly complete overhaul of the user interface AND how our AI Assistant works on the back-end, plus a ton of new, big features, we decided that this is the moment to call this the Assistant 2.0. Let’s go over what’s new, but first a short summary of where we are today!
Nextcloud has a ton of AI features, some big and elaborate, others more basic. Some train a neural network on your data, on your own server to give you a smart inbox in Mail or warnings about unexpected logins from Suspicious Login Detection. Others rely on a pre-trained network that can recognize faces and objects in photos, generate a transcript of your video call or modify text. We can also generate images, translate text and chat messages, and much more. Many of the features are accessible through our easy to use Nextcloud Assistant interface!
Nextcloud Assistant in Text
Many of these features can help you, save you time, or improve your productivity. And best of all: as always we focus on making all these services available of running on your own server. Given even the US Congress doesn’t trust ChatGPT or Microsoft CoPilot, there are good reasons to want control over the data send to and from an AI service! Our Ethical AI Rating system provides you transparency about the AI used to generate your text, image or classify your photos:
Learn more in our blog post about our approach to AI.
Sign up for our webinar and see all the features of Nextcloud Assistant mentioned in this article.
And now, let’s explore what’s new: The two biggest new features are Context Write and Context Chat. These both share a lot in common, both providing AI-assisted outputs for you, but they’re also distinctly different.
Context Chat, in short, is the ability for the Assistant to answer questions about your data. You can ask the Assistant questions about a document you have: Say, you have a PDF with the manual of your new digital watch… you’ll be able to ask the Assistant how to set it up, or how to replace the battery. And if your company has a nicely documented reimbursement process in its documentation in Collectives, you can ask the Assistant questions about this process. This is an absolutely amazing ability!
It is not just limited to your files either: we developed an API so apps like Mail, Calendar, Talk and Deck can make their data available too. You’ll then be able to ask “When do I have my next meeting with my boss?” (from the Calendar) and “What tasks did she ask me to work on as high priority this week?” (from Mail). Or “Give me a summary of the status of project X.” (Deck + chat perhaps?)
Context Write, on the other hand, lets the Assistant rewrite something in a certain requested context, or style. Say, you have a nice poem and want to write another in that same particular style. By providing the original poem, then new text of what you want to say, the Assistant will then create a masterpiece. But you can also give it a form, give it your data, and ask it to provide the text you would need to fill in the form. Impressive, and useful!
Over the last year and a half, we have introduced a wide range of AI features in Nextcloud. Translation, transcription of audio, text and image generation, and more. Then the Nextcloud Assistant came – and we’ve now integrated most of our AI features into the Assistant itself, all available from a unified interface. It is no longer limited to text, but can also generate images and transcribe audio.
Another improvement many of you will appreciate is that you can now see a history, a list of earlier generated images, text or audio transcripts in the Assistant interface. You can even go back and generate more based on your earlier prompts. It is now also possible to cancel a running operation, if desired.
Lastly, as possible inputs to the Assistant you can paste or type text or record audio, and now… select one or more documents as input too!
We also did significant work behind the scenes. We made it possible for administrators to make use of AI models we’ve not yet explicitly integrated into Nextcloud, giving more flexibility and versatility to your choice of models to use.
Especially for larger deployments, we have made it possible to run the LLM and other AI operations like Transcription and Context Chat on a separate server, using our new external app ecosystem for one-click, Docker-based deployments. You can also now use a GPU for text generation, transcription and context chat. This means transcribing an hour-long audio file could, on a CPU, take many minutes while taking only seconds on a GPU!
And, of course, this is not all. We have developed an API which can be used by apps like Deck, Calendar or Mail to support Context Chat. This will allow the Assistant to answer questions about your mail, tasks or upcoming appointments! We look forward to these and other apps implementing support for this feature.
Nextcloud Hub 8 – our next version out very soon – will also feature two more app integrations of the Assistant: First, Nextcloud Talk will be able to show you summaries of conversation in Talk rooms, and Nextcloud Mail will introduce email ‘suggested answers’! If you haven’t yet registered for our Nextcloud Hub 8 Launch Event on April 24, this is the time to do so to find out more!
So now, you ask: When can I get it? We have good news! Nearly all these features have been backported to Nextcloud Hub 7, meaning they are available in the Nextcloud app store right now. All you need to do is update. If you have not yet tried out the assistant, perhaps this is the time to do it!
To run the AI locally, it’s easiest to use LocalAI. A community container exists for our All-in-one, so users of the AIO can simply go to this page to learn more about adding extra containers and then enable the LocalAI container. Be sure to check out our documentation and ask questions in the forums – you could just start below this post!
This brings us to the final piece of the puzzle: For these AI features to work smoothly, you will need some serious hardware. This unfortunately will not be a fun experience on your Raspberry Pi. But! Luckily, a number of businesses have started to provide AI-as-a-Service, running various open source models that Nextcloud supports in a way that you can connect to remotely. While this is great, unfortunately nearly all these companies are American-based, with a few in China. If you live in the USA, that is of course perfect – the data stays in your jurisdiction, and you have a choice of local providers. But for our users in Europe, this is less than ideal. The US jurisdiction unfortunately treats the data of European users with less than ideal respect for privacy, in no small part thanks to the CLOUD act signed into law by the former US president Donald Trump.
But today, we announce that we have been working closely with several major European hosting companies including IONOS, OVHcloud and plusserver to integrate their upcoming AI-as-a-Service solutions. This will make them available to Nextcloud customers and users who wish to take advantage of the advanced AI functionalities in Nextcloud Hub but don’t wish to host AI services themselves. Several Nextcloud customers have expressed interest in these features and are preparing to or in the process of testing them. Here’s what our partners and customers have to say:
OVHcloud is thrilled to collaborate with Nextcloud in delivering state-of-the-art Digital Sovereign AI solutions to our customers. We are aligned with Nextcloud’s commitment to offering secure, privacy-conscious AI functionalities that safeguard training data and adhere strictly to GDPR regulations. We are excited to announce that this integration will become available in Q2 2024, marking a significant milestone in our journey towards responsible AI innovation.
Germain Masse, AI Product Marketing Manager at OVHCloud
We are proud to offer Nextcloud customers a solution that complies with European data protection regulations. Reliable, trustworthy, and local AI services will be key to protecting their digital sovereignty for both the public and private sectors in Europe.
plusKI brings cutting edge, open source AI technologies in a compliant and digitally sovereign service to the German market. We are excited to work with Nextcloud to make this available to their customers as back-end to the new generation of the Nextcloud Assistant.
Christian Schmitz, Director Open Source at plusserver
We at SUNET are happy that that Nextcloud Assistant is 100% Free and Open Source, runs on-premise and provides useful features that we hope will benefit our users going forward. We enjoy the very productive collaboration with Nextcloud.
Mike Nordin, Sunet, the Swedish University Computer Network
In the future, we will make AI services available to colleagues in the state administration directly at the workplace. We want to drive the development of digitally sovereign and open solutions in addition to the use of existing AI services such as ChatGPT. Because, the administration of the future will work in an automated, algorithmized, cloudified and data-based way. To make this vision of the future a reality, we must provide the appropriate tools. This also creates value and jobs in our domestic digital economy.
Minister of Digitisation and Head of the State Chancellery Dirk Schrödter, Schleswig-Holstein, Germany
US technology firms have a large head start on European companies offering large language models and other AI services to smaller firms and end customers. There is a significant number of offerings either using their own models like OpenAI, or providing one or a choice of open models like Llama, Gemma or Mistral as a service. In Europe, Mistral and Aleph Alpha are still working on offering their own models while only a few large hosting providers are looking to offer various open LLM’s in an AI-as-a-service model.
At the same time, there is a growing demand from our customers and users for European AI services. Running AI locally is expensive, and AI-as-a-Service from foreign tech firms is not a solution for most Nextcloud customers in Europe, given the privacy and sovereignty risks inherent to shipping large amounts of crucial data overseas. For this reason, we have been working with a handful of providers to offer their AI services. To speed up the process, we connected them with potential customers. As you can imagine, offering a service when it is unclear who its customers might be is not easy. But several universities are interested in the ability to outsource some heavy AI operations, so there is now movement towards AI services. This means not only our users, but the wider European ecosystem will be able to benefit from this!
Public availability of the European AI services is expected in the course of spring and summer, with some even aiming for late April/early May!
We are super excited about this release and we’d like to thank everybody who contributed, with testing, reviewing code but also the universities who are willing to test the AI services, and of course OVHcloud, IONOS, plusserver and our other partners who’d like to provide these services!
Sign up for our webinar and see all the features of Nextcloud Assistant mentioned in this article.
The post Nextcloud releases Assistant 2.0 and pushes AI-as-a-Service appeared first on Nextcloud.
Ça y est, c’est officiel ! ChatGPT, le célèbre agent conversationnel développé par OpenAI, est désormais accessible à tous sans qu’on ait besoin de se créer un compte. C’est une nouvelle qui devrait ravir les curieux qui souhaitaient tester les capacités de cette intelligence artificielle révolutionnaire sans avoir à s’embêter avec la création d’un énième compte en ligne.
Pour profiter de ChatGPT sans compte, rien de plus simple ! Il vous suffit de vous rendre sur l’application web chat.openai.com ou de télécharger l’application officielle ChatGPT sur votre smartphone, que vous soyez sur iPhone ou Android. Et vous pourrez directement commencer à discuter avec l’IA sans aucune autre formalité.
Par contre, ici en France, ça n’a pas encore l’air actif. J’ai du passer par un VPN via les États-Unis pour en profiter sans avoir à me créer un compte.
Et il faut quand même noter quelques petites limitations par rapport à la version avec compte. Déjà, vous ne pourrez pas sauvegarder ni consulter l’historique de vos conversations, et encore moins les partager avec d’autres utilisateurs. Vous n’aurez pas non plus accès aux conversations vocales ou aux instructions personnalisées. Et surtout, vous serez limité au modèle standard GPT-3.5, comme pour les comptes gratuits. Si vous voulez profiter de la puissance du modèle GPT-4, il faudra alors passer à la caisse et souscrire à l’abonnement payant ChatGPT Plus.
Mais bon, pour une utilisation basique de ChatGPT, la version sans compte est largement suffisante. Vous pourrez poser toutes vos questions, demander des conseils, générer du contenu, et même avoir des conversations à l’infini avec l’IA. Parfait pour découvrir le potentiel de l’IA conversationnelle et vous familiariser avec cet outil fascinant si ce n’est pas encore fait.
D’ailleurs, OpenAI a précisé avoir mis en place des « garde-fous supplémentaires » pour l’utilisation de ChatGPT sans compte. Donc ne soyez pas surpris si l’IA refuse de répondre à certaines de vos questions un peu trop sensibles ou sur des thèmes controversées.
N’oubliez pas que ChatGPT reste un outil imparfait, avec ses limites et ses défauts et qu’il peut parfois se tromper, inventer des choses ou tenir des propos biaisés. Donc gardez toujours votre esprit critique et ne prenez pas tout ce qu’il dit pour argent comptant. Et par pitié, ne l’utilisez pas comme un oracle infaillible ou comme si c’était Wikipédia. Voyez plutôt ça comme un outil permettant de retravailler du texte.
Amusez-vous bien !
Vous avez vu Voice Engine d’OpenAI ? C’est un modèle d’IA qui est capable de générer des voix synthétiques ultra-réalistes à partir d’un simple échantillon audio de 15 secondes. Seulement 15 secondes, oui !
Concrètement, ça veut dire qu’avec cette IA, on peut créer des voix qui ressemblent à s’y méprendre à celles de vraies personnes. Genre on donne un petit extrait de notre voix, et hop, l’IA peut générer un discours entier qui sonne exactement comme nous. C’est à la fois fascinant et un peu flippant, vous trouvez pas ?
OpenAI sont à la pointe de la recherche dans le domaine et ils nous pondent régulièrement des trucs de malade comme Sora. Concernant Voice Engine, ils ont développé la techno fin 2022 et l’ont intégré dans leur API de synthèse vocale ainsi que dans les fonctionnalités vocales de ChatGPT.
Voici les 15 secondes de vraie voix :
Et voici l’audio qui a été généré à partir de ça :
Mais attention, comme un grand pouvoir implique de grandes responsabilités (coucou Peter !), OpenAI joue la carte de la prudence. Ils sont bien conscients que cette technologie pourrait être utilisée à des fins pas très catholiques, genre pour créer des deepfakes audio et induire les gens en erreur. Du coup, ils la déploient pour l’instant à petite échelle, juste auprès de quelques partenaires de confiance.
Et ces partenaires, ils en font quoi de Voice Engine ?
Eh bien figurez-vous qu’ils développent des applications plutôt cools ! Par exemple, Age of Learning l’utilise pour générer des contenus audio éducatifs avec des voix naturelles et expressives. Ou encore HeyGen qui s’en sert pour traduire des vidéos dans différentes langues en conservant la voix du locuteur d’origine. D’ailleurs c’est ce que j’utilise pour ma chaine Youtube en anglais et je peux vous dire que ça coûte une couille. Ça peut aussi aider les personnes non-verbales à communiquer avec une voix unique grâce à Livox. Et même redonner la parole à des patients ayant perdu l’usage de la voix, comme le fait l’institut Norman Prince Neurosciences de Lifespan.
Rassurez-vous, OpenAI a mis en place des garde-fous, comme l’interdiction d’utiliser Voice Engine pour imiter quelqu’un sans son consentement, l’obligation d’obtenir l’accord explicite du locuteur original, ou encore le watermarking des contenus générés pour pouvoir en tracer l’origine. Ils suggèrent également d’abandonner progressivement l’authentification vocale comme mesure de sécurité, mais également d’explorer des réglementations qui permettraient de protéger l’usage des voix dans l’IA, de sensibiliser le public aux deepfakes et de développer des techniques pour tracer l’origine des contenus audio et visuels.
Bref, Voice Engine c’est à la fois excitant et inquiétant. Ce que je vois, c’est que ça ouvre des perspectives folles en termes d’applications, mais ça soulève aussi pas mal de questions sur l’avenir.
Je vous invite à checker l’article d’OpenAI qui détaille leur approche avec plein d’exemples.
This blog post explores the technical and strategic benefits of deploying open-source AI models on Ubuntu. We’ll highlight why it makes sense to use Ubuntu with open-source AI models, and outline the deployment process on Azure.
Authored by Gauthier Jolly, Software Engineer, CPC, and Jehudi Castro-Sierra, Public Cloud Alliance Director, both from Canonical.
Ubuntu Pro elevates the security and compliance aspects of deploying AI models, offering extended security maintenance, comprehensive patching, and automated compliance features that are vital for enterprise-grade applications. Its integration with Confidential VMs on Azure enhances the protection of sensitive data and model integrity, making it an indispensable tool for tasks requiring stringent security measures like ML training, inference, and confidential multi-party data analytics.
Using a public cloud like Azure gives straightforward access to powerful GPUs and Confidential Compute capabilities, essential for intensive AI tasks. These features significantly reduce the time and complexity involved in setting up and running AI models, without compromising on security and privacy. Although some may opt for on-prem deployment due to specific requirements, Azure’s scalable and secure environment offers a compelling argument for cloud-based deployments.
We are going to explore using open models on Azure by creating an instance with Ubuntu, installing NVIDIA drivers for GPU support, and setting up Ollama for running the models. The process is technical, involving CLI commands for creating the resource group, VM, and configuring NVIDIA drivers. Ollama, the chosen tool for running models like Mixtral, is best installed using Snap for a hassle-free experience, encapsulating dependencies and simplifying updates.
Begin by creating a resource group and then a VM with the Ubuntu image using the Azure CLI.
Connexionaz group create --location westus --resource-group ml-workload
az vm create \
--resource-group ml-workload \
--name jammy \
--image Ubuntu2204 \
--generate-ssh-keys \
--size Standard_NC4as_T4_v3 \
--admin-username ubuntu --license-type UBUNTU_PRO
Note the publicIpAddress from the output – you’ll need it to SSH into the VM.
For GPU capabilities, install NVIDIA drivers using Ubuntu’s package management system. Restart the system after installation.
sudo apt update -y sudo apt full-upgrade -y sudo apt install -y ubuntu-drivers-common sudo ubuntu-drivers install sudo systemctl reboot
Important: Standard NVIDIA drivers don’t support vGPUs (fractional GPUs). See instructions on the Azure site for installing GRID drivers, which might involve building an unsigned kernel module (which may be incompatible with Secure Boot).
Snap simplifies the installation of Ollama and its dependencies, ensuring compatibility and streamlined updates. The –beta flag allows you to access the latest features and versions, which might still be under development
sudo snap install --beta ollama
Configure Ollama to use the ephemeral disk
sudo mkdir /mnt/models sudo snap connect ollama:removable-media # to allow the snap to reach /mnt sudo snap set ollama models=/mnt/models
At this point, you can run one of the open models available out of the box, like mixtral or llama2. If you have a fine-tuned version of these models (a process that involves further training on a specific dataset), you can run those as well.
ollama run mixtral
The first run might take a while to download the model.
Now you can use the model through the console interface:
This step is optional, but provides a UI via your web browser.
sudo snap install --beta open-webui
To quickly access the UI without open ports in the Azure security group, you can create an SSH tunnel to your VM using the following command:
ssh -L 8080:localhost:8080 ubuntu@${IP_ADDR}
Go to http://localhost:8080 in your web browser on your local machine (the command above tunnels the traffic from your localhost to the instance on Azure).:
In case you want to make this service public, follow this documentation.
sudo watch -n2 nvidia-smi
Check that the ollama process is using the GPU, you should see something like this:
+---------------------------------------------------------------------------+ | Processes: | | GPU GI CI PID Type Process name GPU Memory | | ID ID Usage | |===========================================================================| | 0 N/A N/A 1063 C /snap/ollama/13/bin/ollama 4882MiB | +---------------------------------------------------------------------------+
Ubuntu’s open-source foundation and robust ecosystem make it a compelling choice for deploying open-source AI models. When combined with Azure’s GPU capabilities and Confidential Compute features, you gain a flexible, secure, and performant AI solution.
Google Next is making its way to Las Vegas, and Ubuntu is joining the journey. As a proud sponsor, Canonical, the publisher of Ubuntu , invites you to join us at the event and visit booth #252 in the Mandalay Bay Expo Hall. As one of the most popular Linux operating systems, Canonical is dedicated to providing commercial support and driving open source innovation across a diverse range of industries and applications. Stop by and learn more about how Canonical and GCP are collaborating to empower businesses with secure and scalable solutions for their cloud computing needs.
Are you an Ubuntu Pro? Test your skills at our booth! Sit down at our workstation and discover any unpatched vulnerabilities on the machine. Showcase your expertise by securing the system completely, and receive exclusive swag as a token of our gratitude.
At Canonical, security is paramount. Ubuntu Pro offers a solution to offload security and compliance concerns for your open source stack, allowing you to concentrate on building and managing your business. Serving as an additional layer of services atop every Ubuntu LTS release, Ubuntu Pro ensures robust protection for your entire software stack, encompassing over 30,000 open source packages. Say farewell to fragmented security measures; Canonical provides a holistic approach, delivering security and support through a unified vendor. Additionally, relish the assurance of vendor-backed SLA support for open source software, providing peace of mind for your operations.
Confidential computing is a revolutionary technology that disrupts the conventional threat model of public clouds. In the past, vulnerabilities within the extensive codebase of the cloud’s privileged system software, including the operating system and hypervisor, posed a constant risk to the confidentiality and integrity of code and data in operation. Likewise, unauthorized access by a cloud administrator could compromise the security of your virtual machine (VM).
Ubuntu Confidential VMs (CVMs) on Google Cloud offer enhanced security for your workloads by utilizing hardware-protected Trusted Execution Environments (TEEs). With the broadest range of CVMs available, Ubuntu enables users on Google Cloud to benefit from the cutting-edge security features of AMD 4th Gen EPYC processors with SEV-SNP and Intel Trust Domain Extensions (Intel TDX).
Empower your organization with Canonical’s AI solutions. We specialize in the automation of machine learning workloads on any environment, whether private or public cloud, or hybrid or multi cloud. We provide an end-to-end MLOps solution to develop and deploy models in a secure, reproducible, and portable manner that seamlessly integrates with your existing technology stack. Let us help you unlock the full potential of AI.
Mark your calendars and make plans to visit Canonical at Google Cloud Next 2024. Whether you’re seeking cutting-edge solutions for cloud computing, robust security measures for your software stack, or innovative AI tools to propel your organization forward, our team will be on hand to offer insights, demonstrations, and personalized consultations to help you harness the power of open source technology for your business. Join us at booth #252 to discover how Canonical and Ubuntu can elevate your digital journey. See you there!
Prompts:
Canonical at Google Next – What you need to know!
Canonical is excited to sponsor Google Cloud Next in Las Vegas, NV April 9-11, 2024.
visit to the Canonical-Ubuntu booth #252 in the Mandalay Bay Expo Hall.
Our team will be available to discuss the following:
While at our booth, earn some awesome swag by showing that you’re an Ubuntu Pro. Take a seat at our workstation to find the unpatched vulnerabilities on the machine! Upgrade the machine to be fully secure to earn awesome swag!
See you at the event
In our previous post, we discussed how to generate Images using Stable Diffusion on AWS. In this post, we will guide you through running LLMs for text generation in your own environment with a GPU-based instance in simple steps, empowering you to create your own solutions.
Text generation, a trending focus in generative AI, facilitates a broad spectrum of language tasks beyond simple question answering. These tasks include content extraction, summary generation, sentiment analysis, text enhancement (including spelling and grammar correction), code generation, and the creation of intelligent applications like chatbots and assistants.
In this tutorial, we will demonstrate how to deploy two prominent large language models (LLM) on a GPU-based EC2 instance on AWS (G4dn) using Ollama, an open source tool for downloading, managing, and serving LLM models. Before getting started, ensure you have completed our technical guide for installing NVIDIA drivers with CUDA on a G4DN instance.
We will utilize Llama2 and Mistral, both strong contenders in the LLM space with open source licenses suitable for this demo.
While we won’t explore the technical details of these models, it is worth noting that Mistral has shown impressive results despite its relatively small size (7 billion parameters fitting into an 8GB VRAM GPU). Conversely, Llama2 provides a range of models for various tasks, all available under open source licenses, making it well-suited for this tutorial.
To experiment with question-answer models similar to ChatGPT, we will utilize the fine-tuned versions optimized for chat or instruction (Mistral-instruct and Llama2-chat), as the base models are primarily designed for text completion.
Let’s get started!
To begin, open an SSH session to your G4DN server and verify the presence of NVIDIA drivers and CUDA by running:
nvidia-smiKeep in mind that you need to have the SSH port open, the key-pair created or assigned to the machine during creation, the external IP of the machine, and software like ssh for Linux or PuTTY for Windows to connect to the server.
If the drivers are not installed, refer to our technical guide on installing NVIDIA drivers with CUDA on a G4DN instance.
Once you have confirmed the GPU drivers and CUDA are set up, proceed to install Ollama. You can opt for a quick installation using their binary, or choose to clone the repository for a manual installation.
To install Ollama quickly, run the following command
curl -fsSL https://ollama.com/install.sh | shLet’s start with Mistral models and view the results by running:
ollama run mistralThis instruction will download the Mistral model (4.1GB) and serve it, providing a prompt for immediate interaction with the model.
Not a bad response for a prompt written in Spanish!. Now let’s experiment with a prompt to write code:
Impressive indeed. The response is not only generated rapidly, but the code also runs flawlessly, with basic error handling and explanations. (Here’s a pro tip: consider asking for code comments, docstrings, and even test functions to be incorporated into the code).
Exit with the /bye command.
Now, let’s enter the same prompt with Llama2.
We can see that there are immediate, notable differences. This may be due to the training data it has encountered, as it defaulted to a playful and informal chat-style response.
Let’s try Llama2 using the same code prompt from above:
The results of this prompt are quite interesting. Following four separate tests, it was clear that the generated responses had not only broken code but also inconsistencies within the responses themselves. It appears that writing code is not one of the out-of-the-box capabilities of Llama2 in this variant (7b parameters, although there are also versions specialized in code like Code-Llama2), but results may vary.
Let’s run a final test with Code-Llama, a Llama model fine-tuned to create and explain code:
We will use the same prompt from above to write the code:
This time, the response is improved, with the code functioning properly and a satisfactory explanation provided.
You now have the option to either continue exploring directly through this interface or start developing apps using the API.
We now have something ready for immediate use. However, for some added fun, let’s install a chat-like web interface to mimic the experience of ChatGPT.
For this test, we are going to use ollama-ui (https://github.com/ollama-ui/ollama-ui).
⚠︎ Please note that this project is no longer being maintained and users should transition to Open WebUI, but for the sake of simplicity, we are going to still use the Ollama-ui front-end.
In your terminal window, clone the ollama-ui repository by entering the following command:
git clone https://github.com/ollama-ui/ollama-uiHere’s a cool trick: when you run Ollama, it creates an API endpoint on port 11434. However, Ollama-ui will run and be accessible on port 8000, thus, we’ll need to ensure both ports are securely accessible from our machine.
Since we are currently running as a development service (without the security features and performance of a production web server), we will establish an SSH tunnel for both ports. This setup will enable us to access these ports exclusively from our local computer with encrypted communication (SSL).
To create the tunnel for both the web-ui and the model’s API, close your current SSH session and open a new one with the following command:
ssh -L 8000:localhost:8000 -L 11434:127.0.0.1:11434 -i myKeyPair.pem ubuntu@<Machine_IP>Once the tunnel is set up, navigate to the ollama-ui directory in a new terminal and run the following command:
cd ollama-ui
makeNext, open your local browser and go to 127.0.0.1:8000 to enjoy the chat web inRunning an LLM model for text generation on Ubuntu on AWS with a GPU instanceterface!
While the interface is simple, it enables dynamic model switching, supports multiple chat sessions, and facilitates interaction beyond reliance on the terminal (aside from tunneling). This offers an alternative method for testing the models and your prompts.
Thanks to Ollama and how simple it is to install the NVIDIA drivers on a GPU-based instance, we got a very straightforward process for running LLMs for text generation in your own environment. Additionally, Ollama facilitates the creation of custom model versions and fine-tuning, which is invaluable for developing and testing LLM-based solutions.
When selecting the appropriate model for your specific use case, it is crucial to evaluate their capabilities based on architectures and the data they have been trained on. Be sure to explore fine-tuned variants such as Llama2 for code, as well as specialized versions tailored for generating Python code.
Lastly, for those aiming to develop production-ready applications, remember to review the model license and plan for scalability, as a single GPU server may not suffice for multiple concurrent users. You may want to explore Amazon Bedrock, which offers easy access to various versions of these models through a simple API call or Canonical MLOps, an end-to-end solution for training and running your own ML models.
The size of the model significantly impacts the production of better results. A larger model is more capable of reproducing better content (since it has a greater capacity to “learn”). Additionally, larger models offer a larger attention window (for “understanding” the context of the question), and allow more tokens as input (your instructions) and output (the response)
As an example, Llama2 offers three main model sizes regarding the parameter number: 7, 13, or 70 billion parameters. The first model requires a GPU with a minimum of 8GB of GPU RAM, whereas the second requires a minimum of 16GB of VRAM.
Let me share a final example:
I will request the 7B parameters version of Llama2 to proofread an incorrect version of this simple Spanish phrase, “¿Hola, cómo estás?”, which translates to “Hi, how are you?” in English.
I conducted numerous tests, all yielding incorrect results like the one displayed in the screenshot (where “óle” is not a valid word, and it erroneously suggests it means “hello”).
Now, let’s test the same example with Llama2 with 13 billion parameters:
While it failed to recognize that I intended to write “hola,” this outcome is significantly better as it added accents, question marks and detected that “ola” wasn’t the right word to use (if you are curious, it means “wave”) .
Google Next is making its way to Las Vegas, and Ubuntu is joining the journey. As a proud sponsor, Canonical, the publisher of Ubuntu , invites you to join us at the event and visit booth #252 in the Mandalay Bay Expo Hall. As one of the most popular Linux operating systems, Canonical is dedicated to providing commercial support and driving open source innovation across a diverse range of industries and applications. Stop by and learn more about how Canonical and GCP are collaborating to empower businesses with secure and scalable solutions for their cloud computing needs.
Are you an Ubuntu Pro? Test your skills at our booth! Sit down at our workstation and discover any unpatched vulnerabilities on the machine. Showcase your expertise by securing the system completely, and receive exclusive swag as a token of our gratitude.
At Canonical, security is paramount. Ubuntu Pro offers a solution to offload security and compliance concerns for your open source stack, allowing you to concentrate on building and managing your business. Serving as an additional layer of services atop every Ubuntu LTS release, Ubuntu Pro ensures robust protection for your entire software stack, encompassing over 30,000 open source packages. Say farewell to fragmented security measures; Canonical provides a holistic approach, delivering security and support through a unified vendor. Additionally, relish the assurance of vendor-backed SLA support for open source software, providing peace of mind for your operations.
Confidential computing is a revolutionary technology that disrupts the conventional threat model of public clouds. In the past, vulnerabilities within the extensive codebase of the cloud’s privileged system software, including the operating system and hypervisor, posed a constant risk to the confidentiality and integrity of code and data in operation. Likewise, unauthorized access by a cloud administrator could compromise the security of your virtual machine (VM).
Ubuntu Confidential VMs (CVMs) on Google Cloud offer enhanced security for your workloads by utilizing hardware-protected Trusted Execution Environments (TEEs). With the broadest range of CVMs available, Ubuntu enables users on Google Cloud to benefit from the cutting-edge security features of AMD 4th Gen EPYC processors with SEV-SNP and Intel Trust Domain Extensions (Intel TDX).
Empower your organization with Canonical’s AI solutions. We specialize in the automation of machine learning workloads on any environment, whether private or public cloud, or hybrid or multi cloud. We provide an end-to-end MLOps solution to develop and deploy models in a secure, reproducible, and portable manner that seamlessly integrates with your existing technology stack. Let us help you unlock the full potential of AI.
Mark your calendars and make plans to visit Canonical at Google Cloud Next 2024. Whether you’re seeking cutting-edge solutions for cloud computing, robust security measures for your software stack, or innovative AI tools to propel your organization forward, our team will be on hand to offer insights, demonstrations, and personalized consultations to help you harness the power of open source technology for your business. Join us at booth #252 to discover how Canonical and Ubuntu can elevate your digital journey. See you there!
Prompts:
Canonical at Google Next – What you need to know!
Canonical is excited to sponsor Google Cloud Next in Las Vegas, NV April 9-11, 2024.
visit to the Canonical-Ubuntu booth #252 in the Mandalay Bay Expo Hall.
Our team will be available to discuss the following:
While at our booth, earn some awesome swag by showing that you’re an Ubuntu Pro. Take a seat at our workstation to find the unpatched vulnerabilities on the machine! Upgrade the machine to be fully secure to earn awesome swag!
See you at the event
J’espère que vous êtes en forme et prêts à en découdre, car aujourd’hui on va parler d’un sujet marrant : GPT-3.5 Turbo d’OpenAI est devenu le nouveau champion toutes catégories de Street Fighter III !
Non, j’ai rien fumé, il y a bien une IA qui a mis la pâtée à tous ses adversaires lors d’un tournoi un peu spécial.
En effet, la semaine dernière, lors du Mistral AI Hackathon à San Francisco, une équipe de passionnés a eu l’idée de génie d’organiser un tournoi un peu particulier. : Faire s’affronter différents modèles de langage sur le cultissime jeu de baston Street Fighter III, pour voir lequel allait sortir vainqueur.
Parce que bon, c’est bien beau de savoir faire la conversation ou générer des images moches, mais quand il s’agit d’envoyer des tatanes dans la tronche, il faut être un peu plus réactif !
Et c’est là que notre pote GPT-3.5 sort les muscles et s’en sort très bien. Contrairement aux algorithmes d’apprentissage par renforcement (deep learning) qui se contentent bêtement d’accumuler des points en fonction de leurs actions, les modèles de langage comme GPT sont capables de comprendre un contexte et d’agir en conséquence.
En gros, ils analysent ce qu’il se passe à l’écran, les mouvements des personnages, leur barre de vie… Et en fonction de ça, ils décident quelle attaque lancer. Un peu comme un joueur humain en fait, sauf qu’eux n’ont pas besoin de café pour rester concentrés.
Les premières bagarres ont opposé différentes versions du modèle Mistral, dans des combats endiablés dignes des plus grands shōnens. Mais très vite, l’équipe a décidé de corser un peu les choses en invitant OpenAI et ses modèles GPT-3.5 et GPT-4 dans l’arène. Et là, mes amis, ça a commencé à sentir le roussi pour la concurrence !
Les poings ont volé, les combos se sont enchaînés, les contres se sont succédés à un rythme infernal. Un vrai feu d’artifice d’uppercuts, de coups spéciaux et de provocations bien senties. Mais au final, après des dizaines de combats acharnés, c’est bien GPT-3.5 (et plus précisément sa dernière version « Turbo ») qui est ressorti vainqueur ! La médaille d’argent revient à Mistral-small-2042, qui a réussi l’exploit de coiffer sur le poteau un modèle GPT-4 en accès anticipé.
Tout ça pour dire que si vous voulez vous mesurer à ces champions, c’est tout à fait possible ! Le code source du projet est disponible sur Github, et vous n’avez même pas besoin d’un supercalculateur pour faire tourner tout ça. Il vous faudra juste dénicher une ROM de jeu de baston 2D ou 3D old school, et le tour est joué. Perso j’ai hâte de voir ce que ça donne sur un bon vieux Tekken 3…
Pour installer et tester LLM Colosseum :
Blague à part, cette expérience montre bien le potentiel hallucinant des modèles de langage pour les jeux vidéo. On peut tout à fait imaginer des PNJ avec lesquels on pourrait interagir de façon totalement naturelle et immersive, des adversaires capables de s’adapter à votre style de jeu et de vous surprendre… Bref, de quoi révolutionner complètement notre façon de jouer ! Après, faudra quand même faire gaffe à pas trop les énerver, on a bien vu ce que ça donnait quand on laissait GPT-3.5 jouer à des wargames… Boum, plus de planète !
Allez, je vous laisse, faut que je retourne taper Zangief moi.
Merci à Lorenper pour l’info et à très vite pour de nouvelles aventures.
