datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
tiny-random-model-summarychronos-wiki-corpus
Chronos Wiki Corpus 📜
540 Wikipedia articles walked into a FAISS index. The retriever said "top 4." The cross-encoder said "just one, actually." The language model said "let me tell you about a Finnish shipwreck." We fixed that last part.
This is the dataset that powers Chronos-3B — a curated, cleaned collection of Wikipedia articles covering the most consequential events, figures, and inventions of the 20th century. Built for RAG. Useful for a lot more than that.
540… See the full description on the dataset page: https://huggingface.co/datasets/TinyModels/chronos-wiki-corpus.TinyModelTokIdsTinyStoriesV2 tokenized with TinyModel tokenizer.
Only stories >= 128 tokens are included and stories are truncated to 128 tokens.
tiny_model_infocodewiz-ai-tiny-model-training-datatiny-interjection-model-alpha
Tiny Interjection Model Alpha Data
Synthetic typed-chat interaction examples generated with qwen3.5-397b-a17b through DigitalOcean Serverless Inference.
The dataset contains 598 training rows and 104 validation rows. Each row includes separate prompt and completion columns for completion-only tuning, plus the structured event stream, floor-control target, and combined text compatibility column.
Intended use
Training and evaluating small interaction models that… See the full description on the dataset page: https://huggingface.co/datasets/jeremysoojk/tiny-interjection-model-alpha.