swiss-ai/africa-sft

--- dataset_info: features: - name: dataset_source dtype: string - name: initial_prompt dtype: string - name: original_metadata dtype: string - name: conversation_branches dtype: string splits: - name: train num_bytes: 43151662 num_examples: 8370 download_size: 21480657 dataset_size: 43151662 configs: - config_name: default data_files: - split: train path: data/train-* license: apache-2.0 --- Placeholder repository for Apertus' african SFT data. The can be found as part of our [SFT mixture](https://huggingface.co/datasets/swiss-ai/apertus-sft-mixture). More details about data provenance, preparation, and statistics can be found in our [tech report](https://github.com/swiss-ai/apertus-tech-report). Sampling, filtering and data-preparation scripts can be found in [our dedicated GitHub repository](https://github.com/swiss-ai/posttrain-data).