ultrachat_200k

← All datasets

Author: hf-mirror · License: mit · Format: parquet

0 downloads · 0 stars · 0 mirrors · 1,624,049,723 · 8 files

UltraChat 200k (HuggingFaceH4): the filtered UltraChat corpus used to train Zephyr-7B — ~200k multi-turn open-ended dialogues generated against GPT-3.5-Turbo across ~30 domains, decontaminated against eval benchmarks. MIT.

Description

UltraChat 200k (HuggingFaceH4): the filtered UltraChat corpus used to train Zephyr-7B — ~200k multi-turn open-ended dialogues generated against GPT-3.5-Turbo across ~30 domains, decontaminated against eval benchmarks. MIT.

Download

Direct download: https://huggingface.co/datasets/HuggingFaceH4/ultrachat_200k/tree/main

Or via the API: curl -L -o file "https://huggingface.co/datasets/HuggingFaceH4/ultrachat_200k/tree/main"

Files

• data/test_gen-00000-of-00001-3d4cd8309148a71f.parquet (80,359,907)
• data/test_sft-00000-of-00001-f7dfac4afe5b93f4.parquet (81,176,553)
• data/train_gen-00000-of-00003-a6c9fb894be3e50b.parquet (243,687,391)
• data/train_gen-00001-of-00003-d6a0402e417f35ca.parquet (243,300,397)
• data/train_gen-00002-of-00003-c0db75b92a2f48fd.parquet (243,497,356)
• data/train_sft-00000-of-00003-a3ecf92756993583.parquet (243,999,189)
• data/train_sft-00001-of-00003-0a1804bcb6ae68c6.parquet (243,897,199)
• data/train_sft-00002-of-00003-ee46ed25cfae92c6.parquet (244,131,731)

Added 2026-09-27. Platform: Distillio — the open source training data collective.