Amélie Viallet

Sheets - a Hugging Face Space by aisheets Create a dataset by describing what you need in a simple text input or by choosing from existing examples. The application will help you design your dataset based on your description.

huggingface.co/spaces/aishe...

Introducing Hugging Face Sheets, where we explore how to create more accurate and reliable structured data with AI and web sources.

Cultivate robust synthetic data and reduce model hallucinations

Exciting news! I’m designing an open-source app that helps AI builders create high-quality datasets in minutes—whether they start with data or not.

🍆 Watch your dataset grow consistently column by column
🪁 Adjust data generation anytime
🦄 Stay in control while automation does the heavy lifting

Art et IA, une cohabitation est-elle possible ? Pas encore de compte ?

shs.cairn.info/dossiers-202...

Cette phrase n’est pas de moi je l'ai lu dans un article de Jean-Lou Fourquet dans son introduction au dossier "Art et IA, une cohabitation est-elle possible ?"

Cela calme la frénésie autour de l'IA qui pourrait s'emparer de tous les sujets, tous les domaines avec pertinence. Elle éclaire des réflexions sur le lien entre art et automatisation de certains “gestes”.

lnkd.in/dPgVApDA

There is a way to make better use of Generative AI.

As a designer working on a synthetic data generator, I attach great importance to working on fundamental “details” that help users consciously use the technology.

Action prevention, flexibility, efficiency, and information transparency.

It echoes a mention I read a few days ago in a post from Dan Shipper summarizing an interview with the CEO of Vercel.

“AI tools are shifting software toward consumption-based billing models, making us capital allocators who decide how much compute the AI consumes.”

From a citizen's perspective, this list of prompts (read each of them and think about it) illustrates well how companies are inciting people to consume AI in an unsustainable way, making it their faithful companion 🐶 all day long.

From a citizen's perspective, this list of prompts (read each of them and think about it) illustrates well how companies are inciting people to consume AI in an unsustainable way, making it their faithful companion 🐶 all day long.

From a UX perspective, I like this design principle of guiding users to achieve their tasks. In this case, the continuous flow of prompt examples does the job.

About the last ChatGPT home page:

“Be more assisted” would be a more accurate title, in my opinion.

Of course L'Express was obliged to use a photo of Sam Altman to best illustrate a conversation with me.
Of course neither a) taking a new picture of me during the interview, nor b) asking me to provide a picture of myself were impossible.
Of course.

Image showing an overview of languages in the FineWeb-c Dataset.

🌍 Big step for multilingual AI data!

The @hf.co community has rated educational content in languages spoken by 1.6 billion people! New additions:
• Japanese
• Italian
• Old High German

These ratings can help enhance training data for major world languages.

Progress bars showing remaining annotations needed for 15 languages in FineWeb-C dataset, ranging from 6 to 593 annotations needed

The finish line is near! We're building FineWeb-Edu for many languages and need your help 🤗

Many FineWeb-C languages are close to 1,000 annotations!

Assamese is 99.4% done, French needs 64 more annotations, Tamil: 216.

Please help us reach the goal: huggingface.co/spaces/data-...

Imagine creating custom datasets and training AI models WITHOUT writing a single line of code. We did and made it a reality.

@hf.co Synthetic Data Generator

Blog: huggingface.co/blog
Space: huggingface.co/spaces/argil...
GitHub: github.com/argilla-io/s...

fra - français - French Join and contribute to the dataset fra - français - French

I've just contributed 20 examples to FineWeb 2 in French! Join me; we are already a couple of annotators there!

data-is-better-together-fineweb-c.hf.space/share-your-p...

In a couple of minutes, we’ll officially make the FineWeb 2 Annotation Sprint.

🎶 Go with your rhythms, and do what you can.
🤏 There is no minimum.
👐 Each contribution is welcomed.

The more we are, the better the result will be.

Vanakkam makkalae , glad that I’ll be leading the FineWeb 2 collaborative annotation sprint for Tamil! 🤗

I’ll be helping to build an open dataset to improve language models for our language. Do join the process of improving models !

huggingface.co/spaces/Huggi...

huggingface.co/spaces/data-...

I am thrilled to see Argilla increasingly used to enable impactful collaborative work around datasets.

Next week, we’ll announce a massive multi-language open annotation sprint to ensure all languages advance equally in AI.

[UX-UI update]
The latest update on the Argilla homepage provides a clear overview of your annotation projects, enhances project monitoring, and highlights the importance of collaboration in data curation.

👀 Who said the Argilla tool was only for text? I am proud of my brilliant teammates for setting up this significant initiative 🤗 @benburtenshaw.bsky.social @davidberenstein.bsky.social @danielvanstrien.bsky.social @dvilasuero.hf.co

Let’s make a generation of amazing open source image generation models from high quality data.

The best image generation models train on human preferences. Unfortunately, many of these datasets are closed. Let’s change that!

🧵 we're building a community dataset and we need help reviewing!

This is great, one of the responses could be "more effort on open datasets"

Soon: Another sprint to label all together a big dataset for many languages! Nominate yourself to lead your language community!

Labelers training AI say they're overworked, underpaid and exploited by big American tech companies Digital workers in Kenya had to sift through horrific online content to train AI, but say they were underpaid, overworked, and got inadequate mental health support. So they're fighting back.

"Naftali was assigned to train AI to recognize and weed out pornography, hate speech and excessive violence, which meant sifting through the worst of the worst content online for hours on end."
So much of AI is based on exploiting workers in precarious conditions 😔
www.cbsnews.com/news/labeler...

What about building a high-quality dataset together and making the result available to the whole Open-Source community?

You are all welcome with or without code skills, with or without a background in AI.

It's about democratizing access to AI projects and making how they work transparent.

⚫️⚪️ If you think transparency is key to building the image generation models of tomorrow, consider contributing to a massive open dataset.

Follow huggingface.co/data-is-bett... not to miss the announcement!

🧧 AIoAI / Curated Images + Sub-Collections | Are.na This curated collection of images, cut-outs, and scrap materials is here to insp…rce or explore the context of any image, we recommend using Google Image Search.

🏺🎞️ Beautiful collection of graphic materials to create better images with AI. www.are.na/aixdesign/ai...