---
license: mit
task_categories:
- question-answering
language:
- ca
pretty_name: Catalan Alpaca GPT4
size_categories:
- 1K<n<10K
---
**Catalan Alpaca-GPT4 Dataset**

A Catalan translation of the Alpaca-GPT4-Spanish dataset, created using DeepSeek Chat V3. Contains 1,980 instruction-response pairs (approximately 2M tokens).

***Dataset Details***

- Source: Spanish Alpaca-GPT4 dataset
- Target Language: Catalan
- Size: 1,980 conversation pairs
- Translation Model: DeepSeek Chat V3

**Purpose**

Released to advance supervised fine-tuning of Large Language Models with Catalan language capabilities.

***Created by***

<a href="http://www.cienciadedades.cat"><img src="https://cdn-uploads.huggingface.co/production/uploads/63d3eeac640bb0f7717701ca/SYTQybHxPPlq1yUuEKIY-.png" width="200" alt="cienciadedades.cat"></a>