Update README.md

README.md

@@ -1,83 +1,66 @@
 ---
 configs:
-  - config_name: minhash_deduped
-    data_files:
-      - split: train
-        path: data/minhash_deduped/*
-  - config_name: consensus
-    data_files:
-      - split: train
-        path: data/consensus/*
-  - config_name: sentence_deduped
-    data_files:
-      - split: train
-        path: data/sentence_deduped/*
+- config_name: minhash_deduped
+  data_files:
+  - split: train
+    path: data/minhash_deduped/*
+- config_name: consensus
+  data_files:
+  - split: train
+    path: data/consensus/*
+- config_name: sentence_deduped
+  data_files:
+  - split: train
+    path: data/sentence_deduped/*
 default: minhash_deduped
-license: apache-2.0
 language:
-  - ar
+- ar
 ---
 
 # AraMix
 
-Arabic pretraining dataset with three subsets:
+AraMix is a deduplicated Arabic pretraining corpus containing 158.8 billion tokens across 167.6 million documents. Rather than scraping the web again, AraMix combines seven publicly available Arabic datasets, applies Arabic-specific quality filtering, and performs cross-dataset deduplication.
 
 ## Subsets
 
-### minhash_deduped
-- **Documents:** 178.9M
-- **Words:** 78.5B
-- **Schema:** `text`, `id`, `source`
-- MinHash-deduplicated Arabic text from 7 source datasets
-
-### consensus
-- **Documents:** 47.9M
-- **Schema:** `text`, `duplicated`, `source`
-- Documents appearing in 2+ independent datasets (high-quality consensus)
-- `duplicated`: number of datasets the document appears in (2-7)
-- `source`: list of dataset names where the document was found
-
-### sentence_deduped
-- **Documents:** 167.6M
-- **Words:** 71.8B
-- **Schema:** `text`, `id`, `source`
-- Sentence-level deduplicated version of minhash_deduped
-- Removes duplicate 3-sentence spans appearing 3+ times
-- 8.1% word reduction from minhash_deduped
+| Subset | Documents | Tokens | Description |
+|--------|-----------|--------|-------------|
+| `sentence_deduped` | 167.6M | 158.8B | MinHash + sentence-level deduplication |
+| `minhash_deduped` | 178.9M | 177.8B | Document-level MinHash deduplication only |
+| `consensus` | 47.9M | 54.1B | Documents appearing in 2+ source datasets |
 
 ## Usage
 
 ```python
 from datasets import load_dataset
 
-# Load minhash_deduped subset (default)
+ds = load_dataset("SultanR/AraMix", "sentence_deduped")
 ds = load_dataset("SultanR/AraMix", "minhash_deduped")
-
-# Load consensus subset
 ds = load_dataset("SultanR/AraMix", "consensus")
-
-# Load sentence_deduped subset
-ds = load_dataset("SultanR/AraMix", "sentence_deduped")
 ```
 
-## Source Datasets
-- HPLT 2.0 (Arabic)
-- CulturaX (Arabic)
-- ArabicWeb24
-- ClusterLab 101B
-- C4 (Arabic)
-- FineWeb-2 (Arabic)
-- FinePDFs (Arabic)
+## Sources
+
+| Source | Tokens | Documents |
+|--------|--------|-----------|
+| CulturaX | 38.4B | 40.8M |
+| ArabicWeb24 | 31.6B | 33.6M |
+| HPLT 2.0 | 30.4B | 33.1M |
+| FineWeb-2 | 24.2B | 30.6M |
+| C4 | 20.4B | 23.0M |
+| ClusterLab 101B | 7.7B | 5.9M |
+| FinePDFs | 6.1B | 648K |
+
+## Pipeline
+
+1. Quality filtering with Arabic-specific thresholds (terminal punctuation, repetition patterns, script ratio)
+2. Document-level MinHash deduplication (5-gram shingles, 14 bands, 8 hashes per bucket)
+3. Sentence-level deduplication (3-sentence spans, minimum 3 occurrences)
+
+## Citation
+
+TBD
 
-## Sentence Deduplication Stats
+## License
 
-| Source | Input Docs | Output Docs | Docs Removed | Words Removed |
-|--------|------------|-------------|--------------|---------------|
-| ArabicWeb24 | 34.5M | 33.6M | 956K | 1.28B |
-| C4 | 26.9M | 23.0M | 3.9M | 510M |
-| ClusterLab 101B | 6.5M | 5.9M | 675K | 668M |
-| CulturaX | 44.2M | 40.8M | 3.4M | 1.15B |
-| FinePDFs | 670K | 648K | 21K | 11M |
-| FineWeb-2 | 32.0M | 30.6M | 1.4M | 1.16B |
-| HPLT 2.0 | 34.0M | 33.1M | 947K | 1.57B |
-| **Total** | **178.9M** | **167.6M** | **11.3M** | **6.34B** |
+See individual source dataset licenses.