Hugging Face's logo

Datasets:
gabtan99
/
pex-conversations

Modalities:
Text
Formats:
json
Sub-tasks:
dialogue-modeling
language-modeling
Languages:
Tagalog
Filipino
Size:
< 1K
Tags:
multi-turn
Libraries:
Datasets
pandas
License:
Dataset card Viewer Files Community
4
gabtan99 commited on
Commit
1c89a90
·
1 Parent(s): cd3d02e

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +7 -1
README.md CHANGED
@@ -24,5 +24,11 @@ task_ids:
24
  # PinoyExchange (PEx) Conversations Dataset
25
 
26
  # Summary
27
- PEx Conversations is a dataset composed of scraped threads from PinoyExchange.com
28
 
 
 
 
 
 
 
 
24
  # PinoyExchange (PEx) Conversations Dataset
25
 
26
  # Summary
27
+ PEx Conversations is a dataset composed of collected threads from PinoyExchange.com (Consisting of Tagalog, English, or Taglish responses)
28
 
29
+ The corpus consists of 45K total scraped conversations from 8 subforums. The data only consists of the user message which means any images, videos, links, or any embdedded html are not collected in the scraping process. All characters have been transliterated to its closest ASCII representation and resolved any unicode errors.
30
+
31
+ # Subforum percentages
32
+ The amount of data per subforum are as follows:
33
+ * Body and Fitness - 3968 conversations, 94395 utterances
34
+ *