| --- |
| license: openrail |
| language: |
| - it |
| tags: |
| - italian, |
| - dataset, |
| - fine-tuning, |
| - SFT, |
| - prompt-response, |
| - json, |
| --- |
| |
| This dataset is a free preview subset of a larger proprietary dataset developed by Egomnia S.p.A. |
|
|
| The version available here represents only a small portion of the full dataset, which is not included in this repository. |
|
|
| The complete dataset can be purchased at: https://ai.egomnia.com/store |
|
|
| This dataset, titled **EmmaSFT 6.0**, consists of a structured JSON collection of prompt–response pairs (over 100,000 samples in the full version), designed for supervised fine-tuning (SFT) of large language models. |
|
|
| It is derived from Egomnia’s internal instruction-following dataset used in the development of **Emma 6**, an open-source conversational model by Egomnia S.p.A. |
|
|
| The dataset is specifically designed for **general-purpose supervised fine-tuning**, with a focus on conversational ability, instruction following, and alignment to user intent. |
|
|
| This dataset is intended for supervised fine-tuning use cases only and is not intended for pretraining. |
|
|
| By downloading and using this dataset (including this preview version), you agree to comply with the OpenRAIL license and the additional terms described below. |
|
|
| The dataset may be used for training, fine-tuning, and evaluating machine learning and artificial intelligence models. Models trained using this dataset may be used freely, including for commercial purposes such as APIs, software products, and other AI-based services. |
|
|
| Redistribution, resale, sublicensing, or sharing of the dataset or any substantial part of it to third parties is strictly prohibited. |
|
|
| Any model, technical documentation, model card, or public description of a system trained using this dataset must include the attribution: “Trained using datasets provided by Egomnia S.p.A.” |
|
|
| The full dataset underlying this preview is part of a proprietary commercial offering by Egomnia S.p.A., an Italian innovative SME and AI research laboratory specializing in high-quality Italian-language datasets for machine learning. |
|
|
| Egomnia datasets are designed to ensure high linguistic accuracy and strong performance in real-world AI applications. |
|
|
| Egomnia strongly supports academic research and the development of artificial intelligence in Italy. For this reason, full versions of the dataset are made available free of charge to universities and accredited research institutions for non-commercial research purposes. |
|
|
| The complete dataset can be purchased at: https://ai.egomnia.com/store |
