Datasets:
Tasks:
Text Generation
Formats:
parquet
Sub-tasks:
language-modeling
Languages:
Danish
Size:
1M - 10M
License:
added notes to readme
Browse files- data/pleias-public-domain/tmp/Danish-PD +1 -0
- paper/paper.md +11 -0
data/pleias-public-domain/tmp/Danish-PD
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
Subproject commit 21a73245f38f5af9db853c12aa5231fc3dca80a3
|
paper/paper.md
CHANGED
@@ -77,6 +77,17 @@ Derivatives models might are thus presented with a unclear legal status. This ma
|
|
77 |
Danish Dynaword seeks to remedy this issue by ...
|
78 |
-->
|
79 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
80 |
## What is a Dynaword
|
81 |
|
82 |
A dynaword is a continously developed dataset resource intended a variety of downstream use-cases within natural language processing. Dynaword does intend to replace existing large scale releases such as fine-web [@fineweb], OSCAR [@OSCAR], or HLPT [@hplt], but rather
|
|
|
77 |
Danish Dynaword seeks to remedy this issue by ...
|
78 |
-->
|
79 |
|
80 |
+
|
81 |
+
<!--
|
82 |
+
Comparable released datasets such as the public datasets by Pleaias (Danish_PD) which consist of >300M words initally seems
|
83 |
+
impressive, but examining the Danish subsection we see that it contains an extensive degree of OCR errors if we apply a simple heuristic alpha ratio filter [cite-gopher] we see that more than 50% of the texts are included.
|
84 |
+
the ratio of words which include only alphabetic characters, we see that 71%
|
85 |
+
|
86 |
+
it 71% of the dataset
|
87 |
+
|
88 |
+
https://huggingface.co/datasets/danish-foundation-models/danish-dynaword/discussions/38
|
89 |
+
-->
|
90 |
+
|
91 |
## What is a Dynaword
|
92 |
|
93 |
A dynaword is a continously developed dataset resource intended a variety of downstream use-cases within natural language processing. Dynaword does intend to replace existing large scale releases such as fine-web [@fineweb], OSCAR [@OSCAR], or HLPT [@hplt], but rather
|