Datasets:
Taemin Lee
commited on
Commit
•
ba76fac
1
Parent(s):
1a4c678
init
Browse files
README.md
CHANGED
@@ -1,3 +1,19 @@
|
|
1 |
-
|
2 |
-
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
### Dataset Summary
|
2 |
+
|
3 |
+
Korean translation of Guanaco via the DeepL API
|
4 |
+
|
5 |
+
Note: There are cases where multilingual data has been converted to monolingual data during batch translation to Korean using the API.
|
6 |
+
|
7 |
+
Below is Guanaco's README.
|
8 |
+
|
9 |
+
----
|
10 |
+
|
11 |
+
This dataset is a subset of the Open Assistant dataset, which you can find here: https://huggingface.co/datasets/OpenAssistant/oasst1/tree/main
|
12 |
+
|
13 |
+
This subset of the data only contains the highest-rated paths in the conversation tree, with a total of 9,846 samples.
|
14 |
+
|
15 |
+
This dataset was used to train Guanaco with QLoRA.
|
16 |
+
|
17 |
+
For further information, please see the original dataset.
|
18 |
+
|
19 |
+
License: Apache 2.0
|
openassistant_best_replies_eval_ko.jsonl
ADDED
The diff for this file is too large to render.
See raw diff
|
|
openassistant_best_replies_train_ko.jsonl
ADDED
The diff for this file is too large to render.
See raw diff
|
|