Datasets:

Modalities:
Text
Formats:
parquet
Languages:
English
ArXiv:
Libraries:
Datasets
Dask
License:
HugoLaurencon HF staff commited on
Commit
f742802
1 Parent(s): ce6e440

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +9 -1
README.md CHANGED
@@ -48,10 +48,18 @@ For example, for the web document `<image_1>text<image_2>`, in `images`, we have
48
 
49
  The images are replaced by their URLs, and the users have to download them themselves, for example with the library `img2dataset`.
50
 
51
- In `metadata`, there is a string that can be transformed into a list with `json.loads(example["metadata"])`. This list will have the same size as the lists of images and texts, and will have a dictionary for each index where there is an image, and a `None` value when there is a text. This dictionary will contain the metadata of the images (original source document, unformatted source, alt-text if present, ...).
52
 
53
  Finally, in `general_metadata`, there is a string that can be transformed into a dictionary, containing the URL of the document, and information about its location in the Common Crawl data.
54
 
 
 
 
 
 
 
 
 
55
  ### Visualization of OBELISC documents
56
 
57
  https://huggingface.co/spaces/HuggingFaceM4/obelisc_visualization
 
48
 
49
  The images are replaced by their URLs, and the users have to download them themselves, for example with the library `img2dataset`.
50
 
51
+ In `metadata`, there is a string that can be transformed into a list with `json.loads(example["metadata"])`. This list will have the same size as the lists of images and texts, and will have a dictionary for each index where there is an image, and a `None` value when there is a text. This dictionary will contain the metadata of the image (original source document, unformatted source, alt-text if present, ...).
52
 
53
  Finally, in `general_metadata`, there is a string that can be transformed into a dictionary, containing the URL of the document, and information about its location in the Common Crawl data.
54
 
55
+ ## Data Splits
56
+
57
+ There is only one split, `train`, that contains 141 million examples.
58
+
59
+ ## Size
60
+
61
+ `OBELISC` with images replaced by their URLs, in arrow format, weights 666.6 GB (unwanted!)
62
+
63
  ### Visualization of OBELISC documents
64
 
65
  https://huggingface.co/spaces/HuggingFaceM4/obelisc_visualization