Commit
•
07e68d7
1
Parent(s):
bab71b5
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,60 +1,7 @@
|
|
| 1 |
---
|
| 2 |
tags:
|
| 3 |
-
-
|
| 4 |
library_name: generic
|
| 5 |
---
|
| 6 |
|
| 7 |
-
#
|
| 8 |
-
Word-vectors created from a large corpus of competitive debate evidence, and data extraction / processing scripts
|
| 9 |
-
|
| 10 |
-
#usage
|
| 11 |
-
```
|
| 12 |
-
import fasttext.util
|
| 13 |
-
ft = fasttext.load_model('debate2vec.bin')
|
| 14 |
-
ft.get_word_vector('dialectics')
|
| 15 |
-
```
|
| 16 |
-
# Download Link
|
| 17 |
-
Github won't let me store large files in their repos.
|
| 18 |
-
* [FastText Vectors Here](https://drive.google.com/file/d/1m-CwPcaIUun4qvg69Hx2gom9dMScuQwS/view?usp=sharing) (~260mb)
|
| 19 |
-
|
| 20 |
-
|
| 21 |
-
# About
|
| 22 |
-
|
| 23 |
-
Created from all publically available Cross Examination Competitive debate evidence posted by the community on [Open Evidence](https://openev.debatecoaches.org/) (From 2013-2020)
|
| 24 |
-
|
| 25 |
-
Search through the original evidence by going to [debate.cards](http://debate.cards/)
|
| 26 |
-
|
| 27 |
-
Stats about this corpus:
|
| 28 |
-
* 222485 unique documents larger than 200 words (DebateSum plus some additional debate docs that weren't well-formed enough for inclusion into DebateSum)
|
| 29 |
-
* 107555 unique words (showing up more than 10 times in the corpus)
|
| 30 |
-
* 101 million total words
|
| 31 |
-
|
| 32 |
-
Stats about debate2vec vectors:
|
| 33 |
-
* 300 dimensions, minimum number of appearances of a word was 10, trained for 100 epochs with lr set to 0.10 using FastText
|
| 34 |
-
* lowercased (will release cased)
|
| 35 |
-
* No subword information
|
| 36 |
-
|
| 37 |
-
The corpus includes the following topics
|
| 38 |
-
|
| 39 |
-
* 2013-2014 Cuba/Mexico/Venezuela Economic Engagement
|
| 40 |
-
* 2014-2015 Oceans
|
| 41 |
-
* 2015-2016 Domestic Surveillance
|
| 42 |
-
* 2016-2017 China
|
| 43 |
-
* 2017-2018 Education
|
| 44 |
-
* 2018-2019 Immigration
|
| 45 |
-
* 2019-2020 Reducing Arms Sales
|
| 46 |
-
|
| 47 |
-
Other topics that this word vector model will handle extremely well
|
| 48 |
-
|
| 49 |
-
* Philosophy (Especially Left-Wing / Post-modernist)
|
| 50 |
-
* Law
|
| 51 |
-
* Government
|
| 52 |
-
* Politics
|
| 53 |
-
|
| 54 |
-
|
| 55 |
-
Initial release is of fasttext vectors without subword information. Future releases will include fine-tuned GPT-2 and other high end models as my GPU compute allows.
|
| 56 |
-
|
| 57 |
-
# Screenshots
|
| 58 |
-
|
| 59 |
-
|
| 60 |
-
|
|
|
|
| 1 |
---
|
| 2 |
tags:
|
| 3 |
+
- text-classification
|
| 4 |
library_name: generic
|
| 5 |
---
|
| 6 |
|
| 7 |
+
# Fasttext nearest neighbors
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|