Datasets:

Languages:
English
Size:
n>1T
ArXiv:
License:
Files changed (5) hide show
  1. .gitignore +1 -1
  2. README.md +32 -5
  3. dolma.py +5 -2
  4. urls/.DS_Store +0 -0
  5. urls/v1_7.txt +0 -0
.gitignore CHANGED
@@ -59,7 +59,7 @@ target/
59
  *.so
60
 
61
  # macOS metadata
62
- .DS_Store
63
 
64
  # ignoring test output
65
  /tests/work/
 
59
  *.so
60
 
61
  # macOS metadata
62
+ *.DS_Store
63
 
64
  # ignoring test output
65
  /tests/work/
README.md CHANGED
@@ -1,6 +1,6 @@
1
  ---
2
  license: odc-by
3
- viewer: true
4
  task_categories:
5
  - text-generation
6
  language:
@@ -28,26 +28,51 @@ More information:
28
 
29
  To learn more about the toolkit used to create Dolma, including how to replicate this dataset, head over our [GitHub project page](https://github.com/allenai/dolma/tree/main/docs)!
30
 
 
 
31
  **2024-04-15: License Change.** We have updated the license of Dolma to [ODC-BY](https://opendatacommons.org/licenses/by/1-0/). Please see this [blog post](https://blog.allenai.org/making-a-switch-dolma-moves-to-odc-by-8f0e73852f44) for more information.
32
 
33
 
34
  ## Versions
35
 
36
- At the moment, there are five versions of Dolma available:
37
 
38
  | **Version** | **Default?** | **Release Date** | **Size** (gzip) | **Description** |
39
  |--|:--:|--|--|--|
40
- | `v1_6` | ✅ | 2024-01-31 | 5.4 TB | The latest version of Dolma, with 3 trillion tokens from a diverse mix of web content, academic publications, code, books, and encyclopedic materials. |
 
41
  | `v1_6-sample` | | 2024-01-31 | 16.4 GB | A smaller sample of Dolma, with roughly 10 billion tokens. Useful for data exploration. |
42
  | `v1_5` | | 2023-10-31 | 6.4 TB | The version of Dolma used to train [OLMo-1B](https://huggingface.co/allenai/OLMo-1B). Roughly 3 trillion tokens. |
43
  | `v1_5-sample` | | 2023-10-31 | 2.9 TB | A sample of roughly 1.9 trillion tokens used to train [OLMo-7B](https://huggingface.co/allenai/OLMo-7B) |
44
  | `v1` | | 2023-08-18 | 6.0 TB | The first version of Dolma. |
45
 
46
- (Size difference between `v1_6` and previous version is due to different set of metadata included in files: we removed redundant metadata in `v1_6`.)
47
 
48
- ## Summary Statistics (v1.6)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
49
 
50
 
 
 
51
  | **Source** | **Doc Type** | **UTF-8 bytes** (GB) | **Documents** (millions) | **Unicode words** (billions) | **Llama tokens** (billions) |
52
  |--|--|--|--|--|--|
53
  | Common Crawl | web pages | 9,022 | 3,370 | 1,775 | 2,281 |
@@ -60,6 +85,8 @@ At the moment, there are five versions of Dolma available:
60
  | **Total** | | **11,519** | **4,367** | **2,318** | **3,059** |
61
 
62
 
 
 
63
 
64
  ## Download
65
 
 
1
  ---
2
  license: odc-by
3
+ viewer: false
4
  task_categories:
5
  - text-generation
6
  language:
 
28
 
29
  To learn more about the toolkit used to create Dolma, including how to replicate this dataset, head over our [GitHub project page](https://github.com/allenai/dolma/tree/main/docs)!
30
 
31
+ **2024-04-17: Dolma v1.7 Release.** We have released an updated version of Dolma that we used to train our latest [OLMo 7B-v1.7](https://huggingface.co/allenai/OLMo-7b-v1.7) model.
32
+
33
  **2024-04-15: License Change.** We have updated the license of Dolma to [ODC-BY](https://opendatacommons.org/licenses/by/1-0/). Please see this [blog post](https://blog.allenai.org/making-a-switch-dolma-moves-to-odc-by-8f0e73852f44) for more information.
34
 
35
 
36
  ## Versions
37
 
38
+ At the moment, there are six versions of Dolma available:
39
 
40
  | **Version** | **Default?** | **Release Date** | **Size** (gzip) | **Description** |
41
  |--|:--:|--|--|--|
42
+ | `v1_7` | ✅ | 2024-04-15 | 4.5 TB | Used to train [OLMo-7B-v1.7](https://huggingface.co/allenai/OLMo-7b-v1.7). |
43
+ | `v1_6` | | 2024-01-31 | 5.4 TB | An update to v1.5 with some bug-fixes. |
44
  | `v1_6-sample` | | 2024-01-31 | 16.4 GB | A smaller sample of Dolma, with roughly 10 billion tokens. Useful for data exploration. |
45
  | `v1_5` | | 2023-10-31 | 6.4 TB | The version of Dolma used to train [OLMo-1B](https://huggingface.co/allenai/OLMo-1B). Roughly 3 trillion tokens. |
46
  | `v1_5-sample` | | 2023-10-31 | 2.9 TB | A sample of roughly 1.9 trillion tokens used to train [OLMo-7B](https://huggingface.co/allenai/OLMo-7B) |
47
  | `v1` | | 2023-08-18 | 6.0 TB | The first version of Dolma. |
48
 
 
49
 
50
+ ## Summary Statistics (v1.7)
51
+
52
+ | **Source** | **Provenance** | **New?** | **Documents** (millions) | **OLMo tokens** (billions) | **Sample Proportion** | **Cutoff Date** | **Processing**
53
+ |--|--|--|--|--|--|--|--|
54
+ | Dolma's CC | [Common Crawl](https://commoncrawl.org/) via Dolma v1.6 | Updated | 875.2 | 1,195.5 | 50% | Mar 2023 | Extracted using the Dolma pipeline; new quality filtering and deduplication steps. |
55
+ | Refined Web | [Refined Web](https://huggingface.co/datasets/tiiuae/falcon-refinedweb) | Yes | 664.0 | 456.4 | 100% | Feb 2023 | Filtered using the Dolma pipeline; new quality filtering and deduplication steps. |
56
+ | StarCoder | [StarCoder](https://huggingface.co/blog/starcoder) | Yes | 206.6 | 263.8 | 100% | May 2023 | No further processing. |
57
+ | C4 | [C4](https://huggingface.co/datasets/c4) via Dolma v1.6 | Updated | 249.9 | 138.4 | 50% | Apr 2019 | Filtered using the Dolma pipeline; new quality filtering and deduplication steps. |
58
+ | Reddit | [PushShift API](https://github.com/pushshift/api) | Updated | 377.4 | 79.9 | 100% | Mar 2023 | Extracted using the Dolma pipeline; new quality filtering and deduplication steps. |
59
+ | Semantic Scholar ([S2ORC](https://aclanthology.org/2020.acl-main.447/) & [S2AG](https://www.semanticscholar.org/product/api)) | [peS2o](https://huggingface.co/datasets/allenai/peS2o) via Dolma v1.6 | No | 38.8 | 57.2 | 100% | Mar 2023 | Same as Dolma v1.6 |
60
+ | arXiv | [RedPajama v1](https://huggingface.co/datasets/togethercomputer/RedPajama-Data-1T) | Yes | 1.5 | 28.0 | 100% | Mar 2023 | No further processing. |
61
+ | StackExchange | [RedPajama v1](https://huggingface.co/datasets/togethercomputer/RedPajama-Data-1T) | Yes | 29.3 | 19.6 | 100% | Mar 2023 | No further processing. |
62
+ | Flan | [Flan](https://arxiv.org/abs/2301.13688) via [Tulu](https://huggingface.co/datasets/allenai/tulu-v2-sft-mixture) | Yes | 52.1 | 16.5 | 100% | Mar 2023 | |
63
+ | CC News | [Common Crawl](https://commoncrawl.org/blog/news-dataset-available) | Yes | 22.0 | 14.3 | 100% | Mar 2023 | Extracted using the Dolma pipeline; new quality filtering and deduplication steps. |
64
+ | OpenWebMath | [OpenWebMath](https://huggingface.co/datasets/open-web-math/open-web-math) via [Proof Pile II](https://huggingface.co/datasets/EleutherAI/proof-pile-2) | Yes | 2.9 | 12.6 | 100% | Oct 2023 | Training subset; no further processing. |
65
+ | Algebraic Stack | [Proof Pile II](https://huggingface.co/datasets/EleutherAI/proof-pile-2) | Yes | 2.8 | 12.6 | 100% | Oct 2023 | Training subset; no further processing. |
66
+ | Project Gutenberg | [Project Gutenberg](https://www.gutenberg.org) via Dolma v1.6 | No | 0.0556 | 5.3 | 100% | Mar 2023 | Same as Dolma v1.6 |
67
+ | MegaWika | [MetaWika](https://huggingface.co/datasets/hltcoe/megawika) | Yes | 3.2 | 4.6 | 100% | Jul 2023 | English web pages cited from Wikipedia; curated using the full Dolma pipeline. |
68
+ | Wikipedia & Wikibooks | [Wikimedia](https://dumps.wikimedia.org) via Dolma v1.6 | No | 6.2 | 3.7 | 200% | Mar 2023 | Same as Dolma v1.6 |
69
+ | **Total** | | | | **2,308.5** | **1,715.1** | | |
70
+
71
+ (A subset of total data was used for training of OLMo 7B-v1.7. The token counts are based on the full dataset, whereas taking into account sampling proportion gives the final actual token counts used for training --- 1.715 trillion tokens.)
72
 
73
 
74
+ ## Summary Statistics (v1.6)
75
+
76
  | **Source** | **Doc Type** | **UTF-8 bytes** (GB) | **Documents** (millions) | **Unicode words** (billions) | **Llama tokens** (billions) |
77
  |--|--|--|--|--|--|
78
  | Common Crawl | web pages | 9,022 | 3,370 | 1,775 | 2,281 |
 
85
  | **Total** | | **11,519** | **4,367** | **2,318** | **3,059** |
86
 
87
 
88
+ (Size difference between `v1_6` and `v1_5` is due to different set of metadata included in files: we removed redundant metadata in `v1_6`.)
89
+
90
 
91
  ## Download
92
 
dolma.py CHANGED
@@ -1,4 +1,4 @@
1
- # Copyright 2023 Allen Institute for AI
2
  #
3
  # Licensed under the Apache License, Version 2.0 (the "License");
4
  # you may not use this file except in compliance with the License.
@@ -35,6 +35,7 @@ _URL_LISTS = {
35
  "v1_5-sample": "urls/v1_5-sample.txt",
36
  "v1_6": "urls/v1_6.txt",
37
  "v1_6-sample": "urls/v1_6-sample.txt",
 
38
  }
39
  _VERSIONS = {
40
  "v1": "1.0.0",
@@ -42,6 +43,7 @@ _VERSIONS = {
42
  "v1_5-sample": "1.5.0",
43
  "v1_6": "1.6.0",
44
  "v1_6-sample": "1.6.0",
 
45
  }
46
  _DATES = {
47
  "v1": "(Aug 2023)",
@@ -49,6 +51,7 @@ _DATES = {
49
  "v1_5-sample": "(Oct 2023)",
50
  "v1_6": "(Jan 2024)",
51
  "v1_6-sample": "(Jan 2024)",
 
52
  }
53
  _BASE_URL = "https://olmo-data.org"
54
 
@@ -84,7 +87,7 @@ class Dolma(datasets.GeneratorBasedBuilder):
84
  for name in _URL_LISTS.keys()
85
  ]
86
 
87
- DEFAULT_CONFIG_NAME = "v1_6"
88
 
89
  def _info(self):
90
  return datasets.DatasetInfo(
 
1
+ # Copyright 2024 Allen Institute for AI
2
  #
3
  # Licensed under the Apache License, Version 2.0 (the "License");
4
  # you may not use this file except in compliance with the License.
 
35
  "v1_5-sample": "urls/v1_5-sample.txt",
36
  "v1_6": "urls/v1_6.txt",
37
  "v1_6-sample": "urls/v1_6-sample.txt",
38
+ "v1_7": "urls/v1_7.txt",
39
  }
40
  _VERSIONS = {
41
  "v1": "1.0.0",
 
43
  "v1_5-sample": "1.5.0",
44
  "v1_6": "1.6.0",
45
  "v1_6-sample": "1.6.0",
46
+ "v1_7": "1.7.0",
47
  }
48
  _DATES = {
49
  "v1": "(Aug 2023)",
 
51
  "v1_5-sample": "(Oct 2023)",
52
  "v1_6": "(Jan 2024)",
53
  "v1_6-sample": "(Jan 2024)",
54
+ "v1_7": "(Apr 2024)",
55
  }
56
  _BASE_URL = "https://olmo-data.org"
57
 
 
87
  for name in _URL_LISTS.keys()
88
  ]
89
 
90
+ DEFAULT_CONFIG_NAME = "v1_7"
91
 
92
  def _info(self):
93
  return datasets.DatasetInfo(
urls/.DS_Store DELETED
Binary file (6.15 kB)
 
urls/v1_7.txt ADDED
The diff for this file is too large to render. See raw diff