Datasets:

Languages:
English
ArXiv:
License:
zwq2018 commited on
Commit
e960cbd
·
verified ·
1 Parent(s): 9622026

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +2 -2
README.md CHANGED
@@ -17,14 +17,14 @@ size_categories:
17
  <img src="./src/logo.png" alt="Image" style="width: 900px;">
18
 
19
 
20
- [![arXiv](https://img.shields.io/badge/arXiv-Paper-<COLOR>.svg)](https://multimodal-textbook.github.io/) [![Project](https://img.shields.io/badge/Project-Website-blue.svg)](https://multimodal-interleaved-textbook.github.io/) [![GitHub](https://img.shields.io/badge/GitHub-Code-181717?logo=github)](https://github.com/DAMO-NLP-SG/multimodal_textbook/tree/master)
21
 
22
 
23
 
24
 
25
  ## Overview
26
 
27
- This dataset is for ["2.5 Years in Class: A Multimodal Textbook for Vision-Language Pretraining"](https://arxiv.org/pdf/2306.07209), containing 6.5M images interleaving with 0.8B text from instructional videos.
28
 
29
  - It contains **pre-training corpus using interleaved image-text format**. Specifically, our multimodal-textbook includes **6.5M keyframes** extracted from instructional videos, interleaving with 0.8B **ASR texts**.
30
  - All the images and text are extracted from online instructional videos (22,000 class hours), covering multiple fundamental subjects, e.g., mathematics, physics, and chemistry.
 
17
  <img src="./src/logo.png" alt="Image" style="width: 900px;">
18
 
19
 
20
+ [![arXiv](https://img.shields.io/badge/arXiv-Paper-<COLOR>.svg)](https://arxiv.org/abs/2501.00958) [![Project](https://img.shields.io/badge/Project-Website-blue.svg)](https://multimodal-interleaved-textbook.github.io/) [![GitHub](https://img.shields.io/badge/GitHub-Code-181717?logo=github)](https://github.com/DAMO-NLP-SG/multimodal_textbook/tree/master)
21
 
22
 
23
 
24
 
25
  ## Overview
26
 
27
+ This dataset is for ["2.5 Years in Class: A Multimodal Textbook for Vision-Language Pretraining"](https://arxiv.org/abs/2501.00958), containing 6.5M images interleaving with 0.8B text from instructional videos.
28
 
29
  - It contains **pre-training corpus using interleaved image-text format**. Specifically, our multimodal-textbook includes **6.5M keyframes** extracted from instructional videos, interleaving with 0.8B **ASR texts**.
30
  - All the images and text are extracted from online instructional videos (22,000 class hours), covering multiple fundamental subjects, e.g., mathematics, physics, and chemistry.