mradermacher/cecilia-2b-instruct-v1-i1-GGUF
2B
•
Updated
•
226
Datasets:
You need to agree to share your contact information to access this dataset
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
By downloading this dataset you agree to the conditions of fair use exposed in the README file.
Log in or Sign Up to review the conditions and access this dataset content.
Maria Silvia Corpus (v1)
This repository contains the v1 version of Maria Silvia, an instruction fine-tunning corpus for Large Language Models focused on Cuban language, culture, and idiosyncracy.
All data is available in three JSON files (train, validation, and test).
The corpus has three configurations:
- compact: Each conversation in full, with one or more turns.
- expanded: Each conversation is expanded to include all prefixes, for example, if a conversation has three turns in the conpact set, you will find three conversations in the expanded set, the first turn, the first two turns, and the three turns.
- salamandra: The expanded set converted to the Salamandra format.
Usage terms
The data in the Maria Silvia datasets is free to use for all purposes. By using it, you waive all rights to sue the authors of the dataset for any damage caused by the use of the dataset, intended or otherwise.
Please beware that some or all data in this dataset has been automatically or semi-automatically generated, and the authors make no claim about its fitness for any particular use.
Download from Hugging Face
You can download the dataset from Hugging Face using the following command:
huggingface-cli download gia-uh/maria-silvia-v1 --local-dir maria-silvia-v1
Or with Python:
from datasets import load_dataset
data = load_dataset("gia-uh/maria-silvia-v1", "compact") # Configuration
License
All data in this folder and subfolders is released with a Creative Commons Zero license.
- Downloads last month
- 5