Hugging Face's logo

Hugging Face TB Research

Enterprise
community
Activity Feed

AI & ML interests

Exploring smol models and high quality web and synthetic datasets, generated by LLMs (TB is for Textbook, as inspired by the "Textbooks are all your need" paper)

Recent Activity

anton-l  new activity about 2 hours ago
HuggingFaceTB/finemath:Create ？
anton-l  new activity 2 days ago
HuggingFaceTB/finemath:[bot] Conversion to Parquet
anton-l  updated a dataset 2 days ago
HuggingFaceTB/math_tasks
View all activity

HuggingFaceTB's activity

anton-l 
in HuggingFaceTB/finemath about 2 hours ago

Create ？
#4 opened about 6 hours ago by
Amyww
merve 
posted an update about 23 hours ago
anton-l 
in HuggingFaceTB/finemath 2 days ago

[bot] Conversion to Parquet
#1 opened 6 days ago by
parquet-converter
anton-l 
updated a dataset 2 days ago
loubnabnl 
in HuggingFaceTB/finemath 2 days ago

Why did you use CC rather than FineWeb to create FineMath?

1
#3 opened 2 days ago by
CryptAL
anton-l 
in HuggingFaceTB/finemath 2 days ago

[Bug] cannot get prompts

1
#2 opened 3 days ago by
BigDong
anton-l 
updated a dataset 2 days ago
Xenova 
in HuggingFaceTB/SmolLM-1.7B 4 days ago

onnx model has additional unknown input

2
#7 opened 4 days ago by
SantoshHF
davanstrien 
posted an update 5 days ago
view post
Post
1512
Introducing FineWeb-C 🌐🎓, a community-built dataset for improving language models in ALL languages.

Inspired by FineWeb-Edu the community is labelling the educational quality of texts for many languages.

318 annotators, 32K+ annotations, 12 languages - and growing! 🌍

data-is-better-together/fineweb-c
loubnabnl 
updated a Space 5 days ago
anton-l 
updated a Space 5 days ago
anton-l 
posted an update 6 days ago
view post
Post
1964
Introducing 📐𝐅𝐢𝐧𝐞𝐌𝐚𝐭𝐡: the best public math pre-training dataset with 50B+ tokens!
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.

HuggingFaceTB/finemath-6763fb8f71b6439b653482c2
merve 
updated a model 6 days ago
merve 
in HuggingFaceTB/SmolVLM-Instruct 6 days ago

Add FT tutorial link
#22 opened 6 days ago by
merve

How to training or fientunee SmolVLM easily?

1
#21 opened 11 days ago by
lucasjin
loubnabnl 
updated a dataset 6 days ago
anton-l 
updated a collection 6 days ago
eliebak 
updated a collection 6 days ago
loubnabnl 
updated a model 6 days ago