Skip to content

Dataset · openbmb

MiniCPM5 SFT Data: Math and Code

openbmb/UltraData-SFT-2605

Over 15M English and Chinese SFT examples for post-training the MiniCPM5-1B-SFT model.

What's inside

UltraData-SFT-2605 is the complete set of core-domain SFT data used during the post-training of MiniCPM5-1B-SFT. It covers math, code, knowledge, instruction following, and other domains, totaling over 15M examples. Languages are English and Chinese; the format is not specified in the available snippet.

What you can do with it

  • Supervised fine-tuning of a language model on math and code
  • Fine-tune an instruction-following model
  • Compare quality of SFT datasets for English and Chinese
Tasks
Text generation, Question answering
Languages
English, Chinese
Size
10M–100M rows
Downloads this month
15.7K
Likes
420
Access
on request on Hugging Face

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("openbmb/UltraData-SFT-2605")