datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
umpalumpas
Compaction Dataset Based on SWE-Smith
This is a sample of a larger dataset. Work in progress.
Parquet Layout
This repository contains a single Parquet file:
File
Rows
What it contains
data/oracle_trajectories.parquet
10
One row per oracle trajectory, preprocessed with linked offline checkpoints, oracle memories, and oracle continuations.
The compaction data is represented by linked offline H2 fixed-milestone
checkpoints:
offline_checkpoint_count… See the full description on the dataset page: https://huggingface.co/datasets/poolside-laguna-hackathon/umpalumpas.py-bug-trace-laguna-xs-2-l1-rolloutspy-bug-trace-laguna-xs-2-l2-rolloutspy-bug-trace-laguna-xs-2-l3-rollouts