Datasets:

Arjun-G-Ravi
/

malayalam-sangraha

Tasks:

Text Generation

Modalities:

Formats:

Languages:

Size:

Libraries:

License:

Dataset card Data Studio Files Files and versions

Arjun-G-Ravi commited on Dec 7, 2025

Commit

2e88356

·

verified ·

1 Parent(s): 1968ce9

Update README.md

Files changed (1) hide show

README.md +3 -1

README.md CHANGED Viewed

@@ -5,9 +5,11 @@ task_categories:
 language:
 - ml
 pretty_name: human-verified-sangraha-malayalam-dataset
 ---
 This is a cleaned version of the malayalam subset of sangraha dataset. This only contains the human verified part of the dataset(which is high quality data obtained from  Indic language PDFs, transcribed data from various Indic language videos, podcasts, movies, courses, etc.)
-The csv dataset is around 32.8 GB. I've also removed the doc_id provided in the dataset, making this ideal for pretraining malayalam LLM.
 For pretraining, I recommend using this dataset along with [Ultimate-malayalam-dataset](https://huggingface.co/datasets/Arjun-G-Ravi/Ultimate-Malayalam-Dataset) for more diverse data.

 language:
 - ml
 pretty_name: human-verified-sangraha-malayalam-dataset
+size_categories:
+- 1M<n<10M
 ---
 This is a cleaned version of the malayalam subset of sangraha dataset. This only contains the human verified part of the dataset(which is high quality data obtained from  Indic language PDFs, transcribed data from various Indic language videos, podcasts, movies, courses, etc.)
+The csv dataset has around 6.3M rows, accounting to 32.8 GB. I've also removed the doc_id provided in the dataset, making this ideal for pretraining malayalam LLM.
 For pretraining, I recommend using this dataset along with [Ultimate-malayalam-dataset](https://huggingface.co/datasets/Arjun-G-Ravi/Ultimate-Malayalam-Dataset) for more diverse data.