Hugging Face's logo

Datasets:
AdaMLLab
/
AraMix

Modalities:
Text
Formats:
parquet
Languages:
Arabic
Size:
100M - 1B
ArXiv:
Libraries:
Datasets
Dask
Polars
Dataset card Data Studio Files
xet
Community
1
SultanR commited on
Commit
afe902d
·
verified ·
1 Parent(s): be72628

Add files using upload-large-folder tool

Browse files
Files changed (49) hide show
  1. README.md +83 -0
  2. data/consensus/train-00001-of-00050.parquet +3 -0
  3. data/consensus/train-00003-of-00050.parquet +3 -0
  4. data/consensus/train-00005-of-00050.parquet +3 -0
  5. data/consensus/train-00008-of-00050.parquet +3 -0
  6. data/consensus/train-00009-of-00050.parquet +3 -0
  7. data/consensus/train-00011-of-00050.parquet +3 -0
  8. data/consensus/train-00012-of-00050.parquet +3 -0
  9. data/consensus/train-00014-of-00050.parquet +3 -0
  10. data/consensus/train-00015-of-00050.parquet +3 -0
  11. data/consensus/train-00017-of-00050.parquet +3 -0
  12. data/consensus/train-00018-of-00050.parquet +3 -0
  13. data/consensus/train-00020-of-00050.parquet +3 -0
  14. data/consensus/train-00022-of-00050.parquet +3 -0
  15. data/consensus/train-00023-of-00050.parquet +3 -0
  16. data/consensus/train-00024-of-00050.parquet +3 -0
  17. data/consensus/train-00025-of-00050.parquet +3 -0
  18. data/consensus/train-00026-of-00050.parquet +3 -0
  19. data/consensus/train-00027-of-00050.parquet +3 -0
  20. data/consensus/train-00029-of-00050.parquet +3 -0
  21. data/consensus/train-00030-of-00050.parquet +3 -0
  22. data/consensus/train-00031-of-00050.parquet +3 -0
  23. data/consensus/train-00033-of-00050.parquet +3 -0
  24. data/consensus/train-00034-of-00050.parquet +3 -0
  25. data/consensus/train-00035-of-00050.parquet +3 -0
  26. data/consensus/train-00036-of-00050.parquet +3 -0
  27. data/consensus/train-00037-of-00050.parquet +3 -0
  28. data/consensus/train-00038-of-00050.parquet +3 -0
  29. data/consensus/train-00041-of-00050.parquet +3 -0
  30. data/consensus/train-00042-of-00050.parquet +3 -0
  31. data/consensus/train-00043-of-00050.parquet +3 -0
  32. data/consensus/train-00044-of-00050.parquet +3 -0
  33. data/consensus/train-00045-of-00050.parquet +3 -0
  34. data/consensus/train-00046-of-00050.parquet +3 -0
  35. data/consensus/train-00047-of-00050.parquet +3 -0
  36. data/minhash_deduped/train-00001-of-00102.parquet +3 -0
  37. data/minhash_deduped/train-00013-of-00102.parquet +3 -0
  38. data/minhash_deduped/train-00021-of-00102.parquet +3 -0
  39. data/minhash_deduped/train-00027-of-00102.parquet +3 -0
  40. data/minhash_deduped/train-00029-of-00102.parquet +3 -0
  41. data/minhash_deduped/train-00030-of-00102.parquet +3 -0
  42. data/minhash_deduped/train-00033-of-00102.parquet +3 -0
  43. data/minhash_deduped/train-00039-of-00102.parquet +3 -0
  44. data/minhash_deduped/train-00040-of-00102.parquet +3 -0
  45. data/minhash_deduped/train-00041-of-00102.parquet +3 -0
  46. data/minhash_deduped/train-00042-of-00102.parquet +3 -0
  47. data/minhash_deduped/train-00057-of-00102.parquet +3 -0
  48. data/minhash_deduped/train-00067-of-00102.parquet +3 -0
  49. data/minhash_deduped/train-00068-of-00102.parquet +3 -0
README.md ADDED
@@ -0,0 +1,83 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ configs:
3
+ - config_name: minhash_deduped
4
+ data_files:
5
+ - split: train
6
+ path: data/minhash_deduped/*
7
+ - config_name: consensus
8
+ data_files:
9
+ - split: train
10
+ path: data/consensus/*
11
+ - config_name: sentence_deduped
12
+ data_files:
13
+ - split: train
14
+ path: data/sentence_deduped/*
15
+ default: minhash_deduped
16
+ license: apache-2.0
17
+ language:
18
+ - ar
19
+ ---
20
+
21
+ # AraMix
22
+
23
+ Arabic pretraining dataset with three subsets:
24
+
25
+ ## Subsets
26
+
27
+ ### minhash_deduped
28
+ - **Documents:** 178.9M
29
+ - **Words:** 78.5B
30
+ - **Schema:** `text`, `id`, `source`
31
+ - MinHash-deduplicated Arabic text from 7 source datasets
32
+
33
+ ### consensus
34
+ - **Documents:** 47.9M
35
+ - **Schema:** `text`, `duplicated`, `source`
36
+ - Documents appearing in 2+ independent datasets (high-quality consensus)
37
+ - `duplicated`: number of datasets the document appears in (2-7)
38
+ - `source`: list of dataset names where the document was found
39
+
40
+ ### sentence_deduped
41
+ - **Documents:** 167.6M
42
+ - **Words:** 71.8B
43
+ - **Schema:** `text`, `id`, `source`
44
+ - Sentence-level deduplicated version of minhash_deduped
45
+ - Removes duplicate 3-sentence spans appearing 3+ times
46
+ - 8.1% word reduction from minhash_deduped
47
+
48
+ ## Usage
49
+
50
+ ```python
51
+ from datasets import load_dataset
52
+
53
+ # Load minhash_deduped subset (default)
54
+ ds = load_dataset("SultanR/AraMix", "minhash_deduped")
55
+
56
+ # Load consensus subset
57
+ ds = load_dataset("SultanR/AraMix", "consensus")
58
+
59
+ # Load sentence_deduped subset
60
+ ds = load_dataset("SultanR/AraMix", "sentence_deduped")
61
+ ```
62
+
63
+ ## Source Datasets
64
+ - HPLT 2.0 (Arabic)
65
+ - CulturaX (Arabic)
66
+ - ArabicWeb24
67
+ - ClusterLab 101B
68
+ - C4 (Arabic)
69
+ - FineWeb-2 (Arabic)
70
+ - FinePDFs (Arabic)
71
+
72
+ ## Sentence Deduplication Stats
73
+
74
+ | Source | Input Docs | Output Docs | Docs Removed | Words Removed |
75
+ |--------|------------|-------------|--------------|---------------|
76
+ | ArabicWeb24 | 34.5M | 33.6M | 956K | 1.28B |
77
+ | C4 | 26.9M | 23.0M | 3.9M | 510M |
78
+ | ClusterLab 101B | 6.5M | 5.9M | 675K | 668M |
79
+ | CulturaX | 44.2M | 40.8M | 3.4M | 1.15B |
80
+ | FinePDFs | 670K | 648K | 21K | 11M |
81
+ | FineWeb-2 | 32.0M | 30.6M | 1.4M | 1.16B |
82
+ | HPLT 2.0 | 34.0M | 33.1M | 947K | 1.57B |
83
+ | **Total** | **178.9M** | **167.6M** | **11.3M** | **6.34B** |
data/consensus/train-00001-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d6d6727c2b8994870c592043617b807090704d5ed8d320382414f0f212cdfc75
3
+ size 1514217449
data/consensus/train-00003-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd087753d48a135e3294ea2edc6897125b080db61b207f4ae8d57e24e9235be6
3
+ size 1507509388
data/consensus/train-00005-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1aa0a8724a0de81297b2f0413e451ecd2d07df5a67c08b7f538114c4b1baa7e0
3
+ size 1502682972
data/consensus/train-00008-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5a6aa07da8c86a36a3593a6349f0e4d4d65f3ffe7d8eec929bf7ef5560cafd77
3
+ size 1483453198
data/consensus/train-00009-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d3a38b19da22e10b6fa7d5dedd3d88ba778915b76094a8ea426c8fdf1019d1f7
3
+ size 1589225105
data/consensus/train-00011-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f0ded3a5cd776a4ab6cd6213df6d710c29626e28b720f34e71ce8cde7b619cf6
3
+ size 1692851092
data/consensus/train-00012-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4c38423c2cb206456f34bf6bbc19b9452420c6925eeec4c7d8ba3c1e5e3b7dc5
3
+ size 1692836989
data/consensus/train-00014-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8df26050041faf15962d97cf2e27f8af36be36b8f72f7c39e6996140c5c045b9
3
+ size 1689496028
data/consensus/train-00015-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:327e5da0413accc080db49ae56795a27ea5060b25030e2edd57b917ebecd9882
3
+ size 1693397536
data/consensus/train-00017-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5bd972fb3b70f5989137cdfcef1567d0c7bbfddf4c9c7ad0546f50c99b499d8f
3
+ size 1649378142
data/consensus/train-00018-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6168abc8b1edc32d03fde2f1bb8044c07ba18cb9c813b55ae81e4d62b7f9fbfd
3
+ size 1643427825
data/consensus/train-00020-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:74aea7e29d688a13c6d1145797f43bee6543235a69b406530cb73f312ee9d6a3
3
+ size 1643565313
data/consensus/train-00022-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a8ecfb69a23a4342174d275989b486bde805d63c164f47db48abc1fcc65a974d
3
+ size 1644242612
data/consensus/train-00023-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:276bd46878f689431b1c9a41512a6dffefc43e09c70313ea62d8364e543384de
3
+ size 1647567236
data/consensus/train-00024-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:048976ff2a6cfaf32426a0af715fcedde5ae561263c681015207dfcaf7a8d279
3
+ size 1646057188
data/consensus/train-00025-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c116c8fd864111e1ec935984fe29a531572dc47db8c3584ad2ace591a929cde8
3
+ size 1648945049
data/consensus/train-00026-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:991a849e0a20c7c6aa697a05c6d70a9313bc27398386f9acb40a4c1c7a85dcc7
3
+ size 1644239899
data/consensus/train-00027-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:28a63dc7766e3ff327279bf90c2c886176ff876006d7623cb25793077ac2c109
3
+ size 1649609104
data/consensus/train-00029-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:58c4cb8b1964772deb4712245e604bf3d54d403e2d80e6c8dee691df8085c9ae
3
+ size 1774351358
data/consensus/train-00030-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f2a22d141851ed9190c602a18accaa0d59376d3dc09a5beb4083572452461317
3
+ size 1895603079
data/consensus/train-00031-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:634c213d2453d29a5d8b1e7085d579550325e4eb2b42621b566573639404e2f5
3
+ size 1895466488
data/consensus/train-00033-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5a49f3fc159e7e878caf1d98c82e1b1c1ee446c5de80b8b0b5393a787fefcdc
3
+ size 1898031583
data/consensus/train-00034-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:96928d12efcda0422b7d28265ff7599cee3269c9b4f74382dee78d975c2eab26
3
+ size 1903494233
data/consensus/train-00035-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6b4a6bab67e6d59ad567292e27365c690d7047587c8a4dd28c0eba3c6588fc73
3
+ size 1901591202
data/consensus/train-00036-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dbd37e9d256968f8425da56e22e42006bf2b62f4359d3f70d2bf934e85356b6d
3
+ size 1900351746
data/consensus/train-00037-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c75c826cfbb4822f9d511157c41b7a32fcab92fb85d03da7f9b976896f698655
3
+ size 1903689262
data/consensus/train-00038-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a328ce1bb66fed712bfbf2a14c6ce5cd76c9edfbbd116c8b08edd4f9d354f87b
3
+ size 1902017098
data/consensus/train-00041-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:72ab65c0baa8fe232e0be80833aff52fa95b4cb547df3ed7b071cfa1a5469db0
3
+ size 1299697104
data/consensus/train-00042-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1be76bea081163b659c21b14eb62d3d164e922e993bca6e5c3b83a2b86459efc
3
+ size 1278637386
data/consensus/train-00043-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:58f29385231922eb47d2dacb06a5fc5e9c9c1bf487eaa78f57780e855951c5d1
3
+ size 1272112492
data/consensus/train-00044-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2cbafaa289deb9dfb01adc665c688753fab7c637df0f9a5ea261e2ef207a3736
3
+ size 1273470417
data/consensus/train-00045-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bdaaf143922625b5b0555a7024257735786fe1b008aaa7aeaa0825cc908b934c
3
+ size 1322464695
data/consensus/train-00046-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6913399d1ef7426bef7c05943594953ace637ac36cc0556b5fb20b25273048dd
3
+ size 1330329080
data/consensus/train-00047-of-00050.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f4d7202bb8111307a236c09f9d9184864ce3fbccb4a7d8510c04263b7f3484a5
3
+ size 1389740821
data/minhash_deduped/train-00001-of-00102.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c6cadb11f2acb9f6cc8406183fbd72d1d58dac956a1b5ebf62506a0c67b975f1
3
+ size 3512308920
data/minhash_deduped/train-00013-of-00102.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c23c7602ed309c9fddd2b8d736db870b3730f0def48ff6846fa68067478375ab
3
+ size 3526425126
data/minhash_deduped/train-00021-of-00102.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:44539e6b85c899ec42b72aa1546519860dd886ee41ee4e1971fbc5d2bfbeffc5
3
+ size 3523948286
data/minhash_deduped/train-00027-of-00102.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:09b2c963f9ed177d2d5a7789bce2517013b7c96fa73b527f6b05cba506120350
3
+ size 3520265601
data/minhash_deduped/train-00029-of-00102.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8b65f44827089f0f586e1ce5f8fb769de00fcdb42f583c180acf37376a0e43c5
3
+ size 3518967125
data/minhash_deduped/train-00030-of-00102.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f727e793d0b17aca4ab29b39ae9ed495afff6e67c1db0dc750c4eb5edfdf49c
3
+ size 3516955269
data/minhash_deduped/train-00033-of-00102.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1bbea360077ed6746dbe117e7ed44b6dd170589dde1fba039c9fbcf81ef74795
3
+ size 3514512385
data/minhash_deduped/train-00039-of-00102.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3097875d8e51a07164ef91d3887b4d151fca1045dc5f1acfb1e941cecd64bd34
3
+ size 3172131408
data/minhash_deduped/train-00040-of-00102.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9627defadfb59da02ed5cc20ab9545b02fc8fc52b7aec64de72c3b37167ccc1c
3
+ size 3033730019
data/minhash_deduped/train-00041-of-00102.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bbc517c9343255192d84fb3d4eb4dfe5c94eeb6c5b23b11b6f9c258246be9543
3
+ size 2947273090
data/minhash_deduped/train-00042-of-00102.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ac998ee2731d25da4153031dfcbc58cb7b703c2c52f665fb66926dd28489eb45
3
+ size 2857067407
data/minhash_deduped/train-00057-of-00102.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:712dffffe92ebc6609f6c86cc7c5985cdd3307389e88e447dae6b7cd4a69d013
3
+ size 3526324317
data/minhash_deduped/train-00067-of-00102.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a1be9bc07a93e435911981220493b26ffeee6f477eca7706c85225ce1a4e0e7f
3
+ size 2624570408
data/minhash_deduped/train-00068-of-00102.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:66f1fe94a5e25c3f98faa92ca2a5d35c587d58ec2fa57f0139defc93cf1ca5c0
3
+ size 2476561751