Skip to content

Dataset · IFM

SFT data for reasoning and instructions

IFM/SFT-Reasoning

Dataset for supervised fine-tuning in instruction following and logical reasoning.

What's inside

The dataset is prepared for supervised fine-tuning and is part of the K2 Horizon collection. It contains text data for training models in instruction following and reasoning, split into parquet partitions with streaming support. The dataset size ranges from 10M to 100M records.

What you can do with it

  • Supervised fine-tuning of language models
  • Training models in logical reasoning
Tasks
Text generation
Size
10M–100M rows
Downloads this month
6.2K
Likes
23

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("IFM/SFT-Reasoning")