-
FineWeb2: One Pipeline to Scale Them All -- Adapting Pre-Training Data Processing to Every Language
Paper • 2506.20920 • Published • 69 -
HuggingFaceFW/fineweb-2
Viewer • Updated • 5.02B • 45.3k • 624 -
68
Scaling FineWeb to 1000+ languages: Step 1: finding signal in 100s of evaluation tasks
📝Evaluate multilingual models using FineTasks
FineData
Enterprise
community
AI & ML interests
We release large pre-training datasets to accelerate open LLM development. Part of the Hugging Face Science team (hf.co/science)
Recent Activity
FineWeb-Edu datasets, classifier and ablation model
-
HuggingFaceFW/fineweb-edu
Viewer • Updated • 3.5B • 98.9k • 740 -
HuggingFaceFW/fineweb-edu-score-2
Viewer • Updated • 13.9B • 6.49k • 78 -
HuggingFaceFW/fineweb-edu-classifier
Text Classification • 0.1B • Updated • 51.6k • • 193 -
HuggingFaceFW/ablation-model-fineweb-edu
Text Generation • 2B • Updated • 1.28k • 16
Ablation models trained for our data experiments.
-
HuggingFaceFW/ablation-exp-textext-warc_trafilatura-28BT
Text Generation • 2B • Updated • 2 -
HuggingFaceFW/ablation-exp-textext-wet-28BT
Text Generation • 2B • Updated • 2 -
HuggingFaceFW/ablation-exp-fw-base_filtering-350BT
Text Generation • 2B • Updated • 2 -
HuggingFaceFW/ablation-exp-dedup-global_minhash-350BT
Text Generation • 2B • Updated • 3
-
1.05k
FineWeb: decanting the web for the finest text data at scale
🍷Generate high-quality web text data for LLM training
-
HuggingFaceFW/fineweb
Viewer • Updated • 52.5B • 247k • 2.33k -
HuggingFaceFW/fineweb-edu
Viewer • Updated • 3.5B • 98.9k • 740 -
HuggingFaceFW/fineweb-edu-score-2
Viewer • Updated • 13.9B • 6.49k • 78
1.8B models trained on 350BT to compare different pretraining datasets
-
HuggingFaceFW/ablation-model-fineweb-edu
Text Generation • 2B • Updated • 1.28k • 16 -
HuggingFaceFW/ablation-model-fineweb-v1
Text Generation • 2B • Updated • 31 • 14 -
HuggingFaceFW/ablation-model-refinedweb
Text Generation • 2B • Updated • 2 • 3 -
HuggingFaceFW/ablation-model-c4
Text Generation • 2B • Updated • 3 • 4
-
FineWeb2: One Pipeline to Scale Them All -- Adapting Pre-Training Data Processing to Every Language
Paper • 2506.20920 • Published • 69 -
HuggingFaceFW/fineweb-2
Viewer • Updated • 5.02B • 45.3k • 624 -
68
Scaling FineWeb to 1000+ languages: Step 1: finding signal in 100s of evaluation tasks
📝Evaluate multilingual models using FineTasks
-
1.05k
FineWeb: decanting the web for the finest text data at scale
🍷Generate high-quality web text data for LLM training
-
HuggingFaceFW/fineweb
Viewer • Updated • 52.5B • 247k • 2.33k -
HuggingFaceFW/fineweb-edu
Viewer • Updated • 3.5B • 98.9k • 740 -
HuggingFaceFW/fineweb-edu-score-2
Viewer • Updated • 13.9B • 6.49k • 78
FineWeb-Edu datasets, classifier and ablation model
-
HuggingFaceFW/fineweb-edu
Viewer • Updated • 3.5B • 98.9k • 740 -
HuggingFaceFW/fineweb-edu-score-2
Viewer • Updated • 13.9B • 6.49k • 78 -
HuggingFaceFW/fineweb-edu-classifier
Text Classification • 0.1B • Updated • 51.6k • • 193 -
HuggingFaceFW/ablation-model-fineweb-edu
Text Generation • 2B • Updated • 1.28k • 16
1.8B models trained on 350BT to compare different pretraining datasets
-
HuggingFaceFW/ablation-model-fineweb-edu
Text Generation • 2B • Updated • 1.28k • 16 -
HuggingFaceFW/ablation-model-fineweb-v1
Text Generation • 2B • Updated • 31 • 14 -
HuggingFaceFW/ablation-model-refinedweb
Text Generation • 2B • Updated • 2 • 3 -
HuggingFaceFW/ablation-model-c4
Text Generation • 2B • Updated • 3 • 4
Ablation models trained for our data experiments.
-
HuggingFaceFW/ablation-exp-textext-warc_trafilatura-28BT
Text Generation • 2B • Updated • 2 -
HuggingFaceFW/ablation-exp-textext-wet-28BT
Text Generation • 2B • Updated • 2 -
HuggingFaceFW/ablation-exp-fw-base_filtering-350BT
Text Generation • 2B • Updated • 2 -
HuggingFaceFW/ablation-exp-dedup-global_minhash-350BT
Text Generation • 2B • Updated • 3