Update README.md
Browse files
README.md
CHANGED
|
@@ -17,6 +17,8 @@ The dataset is ideal to pretrain (or maybe even fine tune) a Large Language Mode
|
|
| 17 |
|
| 18 |
This is the second version of the dataset where I've added some more data and removed a lot of data, which were very short(less than 75 characters). Having a lot off shorter sentences will significantly lower data quality. If you need the older dataset, which contained more than 6 million lines of malayalam text(most of them are super short), see this particular kaggle version: https://www.kaggle.com/datasets/arjungravi/ultimate-malayalam-dataset/versions/3
|
| 19 |
|
|
|
|
|
|
|
| 20 |
# Credits
|
| 21 |
- https://www.kaggle.com/datasets/disisbig/malyalam-news-dataset
|
| 22 |
- https://www.kaggle.com/datasets/parvmodi/english-to-malayalam-machine-translation-dataset?select=train.ml
|
|
|
|
| 17 |
|
| 18 |
This is the second version of the dataset where I've added some more data and removed a lot of data, which were very short(less than 75 characters). Having a lot off shorter sentences will significantly lower data quality. If you need the older dataset, which contained more than 6 million lines of malayalam text(most of them are super short), see this particular kaggle version: https://www.kaggle.com/datasets/arjungravi/ultimate-malayalam-dataset/versions/3
|
| 19 |
|
| 20 |
+
If you want more high quality data on malayalam(presumably for pretraining LLMs), I would recommend mixing this with [malayalam-sangraha](https://huggingface.co/datasets/Arjun-G-Ravi/malayalam-sangraha) dataset(which is more than 32GB of raw malayalam data).
|
| 21 |
+
|
| 22 |
# Credits
|
| 23 |
- https://www.kaggle.com/datasets/disisbig/malyalam-news-dataset
|
| 24 |
- https://www.kaggle.com/datasets/parvmodi/english-to-malayalam-machine-translation-dataset?select=train.ml
|