Gabriel Martín Blázquez's picture

Gabriel Martín Blázquez

gabrielmbmb

·

https://gabrielmb.com

AI & ML interests

ML Engineer

Recent Activity

reacted to anton-l's post with 🚀 4 days ago

Introducing 📐𝐅𝐢𝐧𝐞𝐌𝐚𝐭𝐡: the best public math pre-training dataset with 50B+ tokens! https://huggingface.co./datasets/HuggingFaceTB/finemath Math remains challenging for LLMs and by training on FineMath we see considerable gains over other math datasets, especially on GSM8K and MATH. We build the dataset by: 🛠️ carefully extracting math data from Common Crawl; 🔎 iteratively filtering and recalling high quality math pages using a classifier trained on synthetic annotations to identify math reasoning and deduction. We conducted a series of ablations comparing the performance of Llama-3.2-3B-Base after continued pre-training on FineMath and observe notable gains compared to the baseline model and other public math datasets. We hope this helps advance the performance of LLMs on math and reasoning! 🚀 We’re also releasing all the ablation models as well as the evaluation code. https://huggingface.co./collections/HuggingFaceTB/finemath-6763fb8f71b6439b653482c2

updated a dataset 5 days ago

gabrielmbmb/gsm8k-reasoning-paths-combined

upvoted a paper 5 days ago

Qwen2.5 Technical Report

View all activity

Articles

How we leveraged distilabel to create an Argilla 2.0 Chatbot

Organizations

gabrielmbmb's activity

upvoted a paper 5 days ago

Qwen2.5 Technical Report

Paper • 2412.15115 • Published 6 days ago • 327

upvoted a paper 6 days ago

Smarter, Better, Faster, Longer: A Modern Bidirectional Encoder for Fast, Memory Efficient, and Long Context Finetuning and Inference

Paper • 2412.13663 • Published 7 days ago • 103

upvoted a paper 9 days ago

DotaMath: Decomposition of Thought with Code Assistance and Self-correction for Mathematical Reasoning

Paper • 2407.04078 • Published Jul 4 • 17

upvoted a paper 15 days ago

MAmmoTH-VL: Eliciting Multimodal Reasoning with Instruction Tuning at Scale

Paper • 2412.05237 • Published 19 days ago • 46

upvoted an article 20 days ago

Article

They Said It Couldn’t Be Done

By

•

20 days ago

• 75

upvoted an article 26 days ago

Article

Use Models from the Hugging Face Hub in LM Studio

By

•

27 days ago

• 127

upvoted a collection 29 days ago

Tulu 3 Datasets

All datasets released with Tulu 3 -- state of the art open post-training recipes. • 32 items • Updated 28 days ago • 62

upvoted 2 papers about 1 month ago

Thinking LLMs: General Instruction Following with Thought Generation

Paper • 2410.10630 • Published Oct 14 • 17

Large Language Models Can Self-Improve in Long-context Reasoning

Paper • 2411.08147 • Published Nov 12 • 62

upvoted 2 collections about 2 months ago

Qwen2.5-Coder

Code-specific model series based on Qwen2.5 • 40 items • Updated 28 days ago • 257

SmolLM2

State-of-the-art compact LLMs for on-device applications: 1.7B, 360M, 135M • 15 items • Updated 3 days ago • 195

upvoted an article about 2 months ago

Article

Code a simple RAG from scratch

By

•

Oct 29

• 12

upvoted a collection 2 months ago

🍓 Ichigo v0.3

The experimental family designed to train LLMs to understand sound natively. • 6 items • Updated Nov 11 • 17

upvoted 6 articles 2 months ago

Article

Releasing Outlines-core 0.1.0: structured generation in Rust and Python

Oct 22

• 44

Article

🇮🇹🇯🇵🇧🇷 Generating multilingual instruction datasets with Magpie 🐦‍⬛

By

•

Oct 21

• 18

Article

How to build a custom text classifier without days of human labeling

By

•

Oct 17

• 55

Article

Fixing Gradient Accumulation

Oct 16

• 43

Article

How to optimize your data labelling project with custom interfaces

By

•

Oct 16

• 18

Article

Model2Vec: Distill a Small Fast Model from any Sentence Transformer

By

•

Oct 14

• 61

upvoted a paper 3 months ago

Differential Transformer

Paper • 2410.05258 • Published Oct 7 • 168