Datasets:

Modalities:
Text
Formats:
parquet
Languages:
English
ArXiv:
Libraries:
Datasets
Dask
License:
anas-awadalla commited on
Commit
38d534f
β€’
1 Parent(s): ef62778

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +9 -9
README.md CHANGED
@@ -13,12 +13,12 @@ size_categories:
13
  ---
14
 
15
  <h1 align="center">
16
- <img src="mint-logo.png" width=40px> MINT-1T:<br>Scaling Open-Source Multimodal Data by 10x:<br> A Multimodal Dataset with One Trillion Tokens
17
  </h1>
18
 
19
- MINT-1T is an open-source **M**ultimodal **INT**erleaved dataset with 1 trillion text tokens and 3.4 billion images, a 10x scale-up from existing open-source datasets. Additionally, we include previously untapped sources such as PDFs and ArXiv papers. MINT-1T is designed to facilitate research in multimodal pretraining. MINT-1T is created by a team from the University of Washington in collaboration with Salesforce Research, other academic institutions including Stanford University, University of Texas at Austin, and University of California Berkeley.
20
 
21
- You are currently viewing the HTML subset of MINT-1T. For PDF and ArXiv subsets, please refer to the [MINT-1T collection](https://huggingface.co/collections/mlfoundations/mint-1t-6690216ca4d0df7e518dde1c).
22
 
23
  ## Dataset Details
24
 
@@ -34,21 +34,21 @@ You are currently viewing the HTML subset of MINT-1T. For PDF and ArXiv subsets,
34
 
35
  <!-- This section describes suitable use cases for the dataset. -->
36
 
37
- MINT-1T is designed to facilitate research in multimodal pretraining. The dataset can be used for training multimodal models that can reson about interleaved text and images sequences such as [Idefics2](https://huggingface.co/HuggingFaceM4/idefics2-8b), [XGen-MM](https://huggingface.co/Salesforce/xgen-mm-phi3-mini-instruct-r-v1), and [Chameleon](https://huggingface.co/facebook/chameleon-30b).
38
 
39
  ### Out-of-Scope Use
40
 
41
  <!-- This section addresses misuse, malicious use, and uses that the dataset will not work well for. -->
42
 
43
- MINT-1T was built to make research into large multimodal models more accessible. Using
44
  the dataset to train models that ingest or generate personally identifying information (such
45
- as images of people’s faces and other sensitive content) as well as military applications are all inappropriate use cases of MINT-1T.
46
 
47
  ## Dataset Creation
48
 
49
  ### Curation Rationale
50
 
51
- MINT-1T was created to address a significant gap in the open-source domain by providing a large-scale multimodal interleaved dataset for pre-training large multimodal models. This dataset aims to be a valuable resource for the research community, facilitating open science in multimodal pretraining.
52
 
53
  ### Source Data
54
 
@@ -58,7 +58,7 @@ The dataset is a comprehensive collection of multimodal documents from various s
58
  - PDF documents: Extracted from CommonCrawl WAT dumps covering 2023 to 2024
59
  - ArXiv documents: A subset of papers from the ArXiv repository
60
 
61
- In total, MINT-1T contains 1056.8 million documents, broken down as follows:
62
  - 1029.4 million HTML documents
63
  - 26.8 million PDF documents
64
  - 0.6 million ArXiv documents
@@ -136,7 +136,7 @@ Given these considerations, the following recommendations are provided:
136
  4. Bias Awareness: Researchers and developers should be cognizant of potential biases in the dataset and consider their impact on model training and outputs.
137
 
138
  ## License
139
- We release MINT-1T under a CC-BY-4.0 license, designating it primarily as a research artifact. While the dataset is freely available, users are responsible for ensuring its legal use in commercial settings. Users must independently verify compliance with applicable laws before employing MINT-1T for commercial purposes.
140
 
141
  ## Citation
142
 
 
13
  ---
14
 
15
  <h1 align="center">
16
+ πŸƒ MINT-1T:<br>Scaling Open-Source Multimodal Data by 10x:<br> A Multimodal Dataset with One Trillion Tokens
17
  </h1>
18
 
19
+ πŸƒ MINT-1T is an open-source **M**ultimodal **INT**erleaved dataset with 1 trillion text tokens and 3.4 billion images, a 10x scale-up from existing open-source datasets. Additionally, we include previously untapped sources such as PDFs and ArXiv papers. MINT-1T is designed to facilitate research in multimodal pretraining. MINT-1T is created by a team from the University of Washington in collaboration with Salesforce Research, other academic institutions including Stanford University, University of Texas at Austin, and University of California Berkeley.
20
 
21
+ You are currently viewing the HTML subset of πŸƒ MINT-1T. For PDF and ArXiv subsets, please refer to the [πŸƒ MINT-1T collection](https://huggingface.co/collections/mlfoundations/mint-1t-6690216ca4d0df7e518dde1c).
22
 
23
  ## Dataset Details
24
 
 
34
 
35
  <!-- This section describes suitable use cases for the dataset. -->
36
 
37
+ πŸƒ MINT-1T is designed to facilitate research in multimodal pretraining. The dataset can be used for training multimodal models that can reson about interleaved text and images sequences such as [Idefics2](https://huggingface.co/HuggingFaceM4/idefics2-8b), [XGen-MM](https://huggingface.co/Salesforce/xgen-mm-phi3-mini-instruct-r-v1), and [Chameleon](https://huggingface.co/facebook/chameleon-30b).
38
 
39
  ### Out-of-Scope Use
40
 
41
  <!-- This section addresses misuse, malicious use, and uses that the dataset will not work well for. -->
42
 
43
+ πŸƒ MINT-1T was built to make research into large multimodal models more accessible. Using
44
  the dataset to train models that ingest or generate personally identifying information (such
45
+ as images of people’s faces and other sensitive content) as well as military applications are all inappropriate use cases of πŸƒ MINT-1T.
46
 
47
  ## Dataset Creation
48
 
49
  ### Curation Rationale
50
 
51
+ πŸƒ MINT-1T was created to address a significant gap in the open-source domain by providing a large-scale multimodal interleaved dataset for pre-training large multimodal models. This dataset aims to be a valuable resource for the research community, facilitating open science in multimodal pretraining.
52
 
53
  ### Source Data
54
 
 
58
  - PDF documents: Extracted from CommonCrawl WAT dumps covering 2023 to 2024
59
  - ArXiv documents: A subset of papers from the ArXiv repository
60
 
61
+ In total, πŸƒ MINT-1T contains 1056.8 million documents, broken down as follows:
62
  - 1029.4 million HTML documents
63
  - 26.8 million PDF documents
64
  - 0.6 million ArXiv documents
 
136
  4. Bias Awareness: Researchers and developers should be cognizant of potential biases in the dataset and consider their impact on model training and outputs.
137
 
138
  ## License
139
+ We release πŸƒ MINT-1T under a CC-BY-4.0 license, designating it primarily as a research artifact. While the dataset is freely available, users are responsible for ensuring its legal use in commercial settings. Users must independently verify compliance with applicable laws before employing MINT-1T for commercial purposes.
140
 
141
  ## Citation
142