Datasets:

Modalities:
Text
Formats:
parquet
ArXiv:
Libraries:
Datasets
Dask
License:
xiaotianhan commited on
Commit
3efcc8d
1 Parent(s): aae309d

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +21 -1
README.md CHANGED
@@ -34,13 +34,29 @@ The dataset is organized in a format that captures both text and images in their
34
 
35
  ```json
36
  {
37
- "URL": "...", # The URL of the source document.
38
  "text_list": [...], # List of extracted text segments, None if the element is an image.
39
  "image_list": [...], # List of image URLs, None if the element is a text segment.
40
  "metadata": {...} # Metadata containing information about the extraction process (e.g., processing details, timestamps).
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
41
  }
42
  ```
43
 
 
44
  ### Interleaved Text and Images
45
 
46
  The **text_list** and **image_list** are designed as parallel arrays, maintaining the sequence of the document. This interleaving structure allows models to reconstruct the flow of the original document:
@@ -66,6 +82,10 @@ The **InfiMM-WebMath-40B** dataset was created through a comprehensive multi-sta
66
  1. **Base Text Download**: The dataset is available for download as a set of web documents with interleaved text and image URLs.
67
  2. **Image Download**: Users need to download images according to the image URLs provided.
68
 
 
 
 
 
69
  # License
70
 
71
  **InfiMM-WebMath-40B** is made available under an ODC-By 1.0 license; users should also abide by the CommonCrawl ToU: [https://commoncrawl.org/terms-of-use/](https://commoncrawl.org/terms-of-use/). We do not alter the license of any of the underlying data.
 
34
 
35
  ```json
36
  {
37
+ "URL": "...", # The URL of the source document.
38
  "text_list": [...], # List of extracted text segments, None if the element is an image.
39
  "image_list": [...], # List of image URLs, None if the element is a text segment.
40
  "metadata": {...} # Metadata containing information about the extraction process (e.g., processing details, timestamps).
41
+ "metadata": { # Metadata containing information about the extraction process (e.g., processing details, timestamps).
42
+ "ft_lang_label", # Type of languages detected by fastText
43
+ "ft_lang_prob", # Probability of type of language detected by fastText
44
+ "math_prob", # First round math content detection with high recal fastText model
45
+ "size",
46
+ "snap", # Timestamp of Common Crawl snapshot
47
+ "text_gpt3_token_len",
48
+ "char_repetition_ratio",
49
+ "word_repetition_ratio",
50
+ "special_character_ratio",
51
+ "punctuation_ratio",
52
+ "nsfw_num_words", # Number of words which are NSFW
53
+ "has_unicode_error", # If there's any unicode error exists
54
+ "math_prob_llama3", # Probability of second round math detection with high precision fastText model
55
+ }
56
  }
57
  ```
58
 
59
+
60
  ### Interleaved Text and Images
61
 
62
  The **text_list** and **image_list** are designed as parallel arrays, maintaining the sequence of the document. This interleaving structure allows models to reconstruct the flow of the original document:
 
82
  1. **Base Text Download**: The dataset is available for download as a set of web documents with interleaved text and image URLs.
83
  2. **Image Download**: Users need to download images according to the image URLs provided.
84
 
85
+ ### Note
86
+
87
+ If you want more data with more precision, you can always use higher thresholds with `math_prob` and `math_prob_llama3` fields in `metadata`.
88
+
89
  # License
90
 
91
  **InfiMM-WebMath-40B** is made available under an ODC-By 1.0 license; users should also abide by the CommonCrawl ToU: [https://commoncrawl.org/terms-of-use/](https://commoncrawl.org/terms-of-use/). We do not alter the license of any of the underlying data.