This dataset contains parallel instructions from teknium/OpenHermes-2.5 and Iker/OpenHermes-2.5-Spanish.
The Open Hermes 2.5 dataset includes 1 million primarily synthetically generated instruction and chat samples.
This is a translation of the Open Hermes 2.5 dataset to Spanish using the Iker/TowerInstruct-13B-v0.1-EN2ES model. For more information about how the dataset was compiled, refer to the Dataset Card.… See the full description on the dataset page:
https://huggingface.co/datasets/Iker/OpenHermes-2.5-English-Spanish.