Hugging Face's logo

Datasets:
TeraflopAI
/
SEC-EDGAR

Tasks:
Text Generation
Text Classification
Modalities:
Text
Languages:
English
Size:
1M - 10M
Tags:
finance
edgar
sec
License:
Dataset card Data Studio Files
xet
 Community
2
conceptofmind commited on
Commit
200be8d
·
verified ·
1 Parent(s): bdc0b8f

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -0
README.md CHANGED
@@ -43,6 +43,7 @@ Samples per document type:
43
  | Total | 8,055,455 |
44
 
45
  To collect the total token counts of each filing, we used the [Comma v0.1 tokenizer](https://huggingface.co/common-pile/comma-v0.1-1t), a BPE-based tokenizer with a vocabulary size of 64,000. The dataset encompasses a total of 43 billion clean tokens for training LLMs and building retrieval pipelines.
 
46
 
47
  Total token counts for each filing:
48
 
 
43
  | Total | 8,055,455 |
44
 
45
  To collect the total token counts of each filing, we used the [Comma v0.1 tokenizer](https://huggingface.co/common-pile/comma-v0.1-1t), a BPE-based tokenizer with a vocabulary size of 64,000. The dataset encompasses a total of 43 billion clean tokens for training LLMs and building retrieval pipelines.
46
+ ![Total tokens chart](total_tokens_per_filing_millions.png)
47
 
48
  Total token counts for each filing:
49