Gabriel Martín Blázquez
@gabrielmb.com
1.8K followers 140 following 12 posts
ML Engineer @hf.co 🤗 Building tools for you to take care of your datasets like Argilla or distilabel!
Posts Media Videos Starter Packs
gabrielmb.com
SmolLM2 paper is out! We wrote a paper detailing the steps we took to train one of the best smol LM 🤏 out there: pre-training and post-training data, training ablations and some interesting findings 💡

Go check it out and don't hesitate to write your thoughts/questions in the comments section!
gabrielmb.com
distilabel ⚗️ reached the 2k ⭐️ on GitHub!
Reposted by Gabriel Martín Blázquez
lewtun.bsky.social
We are reproducing the full DeepSeek R1 data and training pipeline so everybody can use their recipe. Instead of doing it in secret we can do it together in the open!

Follow along: github.com/huggingface/...
GitHub - huggingface/open-r1: Fully open reproduction of DeepSeek-R1
Fully open reproduction of DeepSeek-R1. Contribute to huggingface/open-r1 development by creating an account on GitHub.
github.com
Reposted by Gabriel Martín Blázquez
anton-l.bsky.social
Introducing 📐FineMath: the best open math pre-training dataset with 50B+ tokens!

Math remains challenging for LLMs and by training on FineMath we see considerable gains over other math datasets, especially on GSM8K and MATH.

🤗 huggingface.co/datasets/Hug...

Here’s a breakdown 🧵
A plot showing increased performance of Llama-3.2-3B when pretrained on FineMath
Reposted by Gabriel Martín Blázquez
jfcalvo.hf.co
🚀 Argilla v2.6.0 is here! 🎉

Let me show you how EASY it is to export your annotated datasets from Argilla to the Hugging Face Hub. 🤩

Take a look to this quick demo 👇

💁‍♂️ More info about the release at github.com/argilla-io/a...

#AI #MachineLearning #OpenSource #DataScience #HuggingFace #Argilla
gabrielmb.com
That's 100% true. To be honest, all the regular expressions that I've used in the last months have been written by an LLM... Most of the time they work at first try, but when they don't it's a pain.
gabrielmb.com
How many regular expressions have you written without the help of an LLM since ChatGPT appeared?
Reposted by Gabriel Martín Blázquez
thomwolf.bsky.social
The FineWeb team is happy to finally release "FineWeb2" 🥂🥳

FineWeb 2 extends the data driven approach to pre-training dataset design that was introduced in FineWeb 1 to now covers 1893 languages/scripts

Details: huggingface.co/datasets/Hug...

A detailed open-science tech report is coming soon
Reposted by Gabriel Martín Blázquez
benburtenshaw.bsky.social
For anyone interested in fine-tuning or aligning LLMs, I’m running this free and open course called smol course. It’s not a big deal, it’s just smol.

🧵>>
gabrielmb.com
It's just me or the latest Claude 3.5 Sonnet is too prone to generate code when asking technical questions not directly related to coding?
Reposted by Gabriel Martín Blázquez
andimara.bsky.social
Let's go! We are releasing SmolVLM, a smol 2B VLM built for on-device inference that outperforms all models at similar GPU RAM usage and tokens throughputs.

SmolVLM can be fine-tuned on a Google collab and be run on a laptop! Or process millions of documents with a consumer GPU!
gabrielmb.com
As part of the SmolTalk release, the dataset mixture used for @huggingface.bsky.social SmolLM2 model, we built a new version of the MagPie Ultra dataset using Llama 405B Instruct.

It contains 1M rows of multi-turn conversations with diverse instructions!

huggingface.co/datasets/arg...
argilla/magpie-ultra-v1.0 · Datasets at Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
huggingface.co
Reposted by Gabriel Martín Blázquez
dvilasuero.hf.co
Let's make AI more inclusive.

At @huggingface.bsky.social we'll launch a huge community sprint soon to build high-quality training datasets for many languages.

We're looking for Language Leads to help with outreach.

Find your language and nominate yourself:
forms.gle/iAJVauUQ3FN8...
Reposted by Gabriel Martín Blázquez
dvilasuero.hf.co
I am very excited to launch a new community initiative next week.

Let's build the largest open community dataset to evaluate and improve image generation models.

Follow:
huggingface.co/data-is-bett...

And stay tuned here
data-is-better-together (Data Is Better Together)
Building better datasets together
huggingface.co
Reposted by Gabriel Martín Blázquez
loubnabnl.hf.co
Making SmolLM2 more reproducible: open-sourcing our training & evaluation toolkit 🛠️ github.com/huggingface/...

Pre-training & evaluation code, synthetic data generation pipelines, post-training scripts, on-device tools & demos

Apache 2.0. V2 data mix coming soon!

Which tools should we add next?
GitHub - huggingface/smollm: Everything about the SmolLM & SmolLM2 family of models
Everything about the SmolLM & SmolLM2 family of models - GitHub - huggingface/smollm: Everything about the SmolLM & SmolLM2 family of models
github.com
gabrielmb.com
We will soon release all the distilabel code used to generate the datasets. As a sneak peak, you can already check the code used for MagPie Ultra v1.0 here:

github.com/huggingface/...
smollm/distilabel_pipelines at main · huggingface/smollm
Contribute to huggingface/smollm development by creating an account on GitHub.
github.com
gabrielmb.com
The dataset allowed to enhance the instruction following and reasoning of SmolLM2 with respect to the previous version. It also includes instructions for rewriting, summarization and function calling.
gabrielmb.com
Excited to announce the SFT dataset used for @huggingface.bsky.social SmolLM2!

The dataset for SmolLM2 was created by combining multiple existing datasets and generating new synthetic datasets, including MagPie Ultra v1.0, using distilabel.

Check out the dataset:
huggingface.co/datasets/Hug...
HuggingFaceTB/smoltalk · Datasets at Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
huggingface.co
Reposted by Gabriel Martín Blázquez
benburtenshaw.bsky.social
Here's a notebook where I do SFT SmolLM2 on the synthetic dataset: colab.research.google.com/drive/1lioed...

thanks @philschmid.bsky.social for the finetuning code
thanks @huggingface.bsky.social for the smol model
thanks @qgallouedec.bsky.social and friends for TRL
Google Colab
colab.research.google.com
gabrielmb.com
The great exile!

For those who don’t know me, I’m Gabriel, ML Engineer at @huggingface.bsky.social where I work developing tools like distilabel or Argilla for you to take care of your data 🤗

The content of my posts here will be mainly related to synthetic data and LLM post-training.