datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
SAND-Post-Training-Dataset
SAND-Post-Training-Dataset: High-Quality Synthetic Reasoning Dataset Built with AMD GPUs
Dataset Summary
We introduce the SAND-Post-Training-Dataset, a high-quality synthetic reasoning dataset for mathematics and science built entirely using a synthetic data pipeline running on the AMD ROCm™ stack and AMD Instinct™ MI325 GPUs.
This dataset prioritizes difficulty and novelty over volume, demonstrating that high-difficulty synthetic data can elevate… See the full description on the dataset page: https://huggingface.co/datasets/amd/SAND-Post-Training-Dataset.dqs-post-training
DQS Post-Training Preference Data
Strict English-to-Korean preference data for three post-training objectives.
All three configurations contain the same ordered set of 5,200 preference
examples after source-quality review and exclusion of one Teacher/Student pair
with no response-level preference.
Run-prefixed layout
The original root-level mpo/, cpo/, dpo/, and manifest.json are retained as the legacy Gemma release for compatibility with existing download… See the full description on the dataset page: https://huggingface.co/datasets/alwaysgood/dqs-post-training.