datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
breaking-bad-volume-constrained
Breaking Bad Volume-Constrained
61,887 fracture assemblies from 696 source objects,
with 265,092 embedded fragment meshes. This release contains the official
volume-constrained everyday and artifact subsets. It is separate from the
original Breaking Bad version and does not include other or GARF point clouds.
from datasets import load_dataset
objects = load_dataset("AssemblyWorld/breaking-bad-volume-constrained", split="all")
sample = objects[0]
parts = sample["parts"]
Pin… See the full description on the dataset page: https://huggingface.co/datasets/AssemblyWorld/breaking-bad-volume-constrained.Ordering_ConstrainedThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos",
"gripper.pos"
],
"shape": [
6… See the full description on the dataset page: https://huggingface.co/datasets/justintiensmith/Ordering_Constrained.scaling-data-constrained-llms
Scaling Data-Constrained Language Models with Synthetic Data
This repository provides the pre-training corpora used in Scaling Data-Constrained Language Models with Synthetic Data (Findings of EACL 2026).
Overview
This repository contains multiple corpora designed to study data augmentation strategies for pre-training Japanese LLMs under a data-constrained data setting.
Starting from a limited Japanese Web corpus and a larger English Web corpus, we construct three… See the full description on the dataset page: https://huggingface.co/datasets/llm-jp/scaling-data-constrained-llms.Ordering_Constrained_No_DepthThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos",
"gripper.pos"
],
"shape": [
6… See the full description on the dataset page: https://huggingface.co/datasets/justintiensmith/Ordering_Constrained_No_Depth.Ordering_Constrained_ParaphrasesThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos",
"gripper.pos"
],
"shape": [
6… See the full description on the dataset page: https://huggingface.co/datasets/justintiensmith/Ordering_Constrained_Paraphrases.Constrained-IR-Benchmark
Constrained-IR-Benchmark
10-domain retrieval test set. Each domain is a separate config with a test split.
Domains
config
samples
business_selection
25
company
25
course_selection
25
drug
25
enterprise_collaboration
25
phone
25
rental_house
25
restaurant
25
resume
25
vector_database
25
Load
from datasets import load_dataset
# 单个 domain
ds = load_dataset("WangSenQWQ/Constrained-IR-Benchmark", "phone", split="test")… See the full description on the dataset page: https://huggingface.co/datasets/WangSenQWQ/Constrained-IR-Benchmark.Constrained_Indic_Codemixingconstrained_language
Dataset Card for constrained_language (pre-training data for simplified English)
Dataset Summary
This dataset is one of the two datasets published by "Honey, I Shrunk the Language: Language Model Behavior at Reduced Scale" (https://arxiv.org/abs/2305.17266).
The dataset available at this link is the pre-training data constrained by vocabulary. The other published data i.e. the pre-training data that is not constrained by vocabulary is available at… See the full description on the dataset page: https://huggingface.co/datasets/text-machine-lab/constrained_language.constrained-instruction-training-pool
Constrained instruction training pool
Public prompts for writing tasks, many of them carrying a constraint a program can check, from ten
datasets read at the pinned revisions named below and one layer built here from them. The pool is
laid out twice. Train on either layer or on both.
pool.jsonl
Every source rewritten into one shape, 462652 rows, one JSON object per line, with these fields.
Field
What it holds
id
a row identifier unique within this file… See the full description on the dataset page: https://huggingface.co/datasets/Emulated-Inc/constrained-instruction-training-pool.exp10-constrained-choice-steeringOrdering_Constrained_No_Depth_50_demosThis dataset was created using LeRobot.
Dataset Description
This is a 50-episode subset of
justintiensmith/Ordering_Constrained_No_Depth.
It contains 12, 13, 12, and 13 demonstrations for the first, second, third, and fourth
positions from the left, respectively. Episodes were selected evenly across each source
25-episode group, then reindexed to 0–49. All five camera videos were trimmed to include
only the selected demonstrations.
Selected source episode IDs:
First… See the full description on the dataset page: https://huggingface.co/datasets/justintiensmith/Ordering_Constrained_No_Depth_50_demos.pick-place-yellow_cube_constrainedThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so_follower",
"total_episodes": 55,
"total_frames": 26351,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:55"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/XixinZhang/pick-place-yellow_cube_constrained.Ordering_Constrained_OLDThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos",
"gripper.pos"
],
"shape": [
6… See the full description on the dataset page: https://huggingface.co/datasets/justintiensmith/Ordering_Constrained_OLD.exp-syh-r2egym-askllm-constrained_glm_4.7_traces_jupiterOrdering_Constrained_Black_BinThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos",
"gripper.pos"
],
"shape": [
6… See the full description on the dataset page: https://huggingface.co/datasets/justintiensmith/Ordering_Constrained_Black_Bin.instruction-following-rl-content-constrained-30k
Instruction-Following RL Content-Constrained 30K
Dataset summary
Instruction-Following RL Content-Constrained 30K is a training dataset for precise instruction following and reinforcement learning from verifiable rewards (RLVR). The current cleaned revision contains 29,520 heterogeneous, single-turn user prompts. Each prompt combines a substantive task with one to five explicit output constraints, such as keyword inclusion or exclusion, response length… See the full description on the dataset page: https://huggingface.co/datasets/wflying/instruction-following-rl-content-constrained-30k.Ordering_Constrained_Black_Bin_SequencingThis dataset was created using LeRobot.
Dataset Description
This is a sequencing-prompt variant of
justintiensmith/Ordering_Constrained_Black_Bin.
The observations, actions, episode ordering, and videos are unchanged. Only the task prompts and
their task indices have been rewritten.
The black bin is to the left of the four-item sequence. The dataset uses five paraphrases for each
of these relations: first from the left, second from the left, second from the right, and… See the full description on the dataset page: https://huggingface.co/datasets/justintiensmith/Ordering_Constrained_Black_Bin_Sequencing.exp-syh-r2egym-askllm-constrained_glm_4.7_traces_jupiter_cleanedpick-place-yellow_cube_constrained_leftThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so_follower",
"total_episodes": 61,
"total_frames": 25600,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:61"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/XixinZhang/pick-place-yellow_cube_constrained_left.Ordering_Constrained_Black_Bin_Sequencing_RobometerThis dataset was created using LeRobot.
Dataset Description
This is a sequencing-prompt variant of
justintiensmith/Ordering_Constrained_Black_Bin.
The observations, actions, episode ordering, and videos are unchanged. Only the task prompts and
their task indices have been rewritten.
The black bin is to the left of the four-item sequence. The dataset uses five paraphrases for each
of these relations: first from the left, second from the left, second from the right, and… See the full description on the dataset page: https://huggingface.co/datasets/justintiensmith/Ordering_Constrained_Black_Bin_Sequencing_Robometer.repro-learning-fingerprints-for-medical-time-series-with-redundancy-constrained-info-traces
Agent traces
Agent sessions published from a Trackio Logbook.
pick-place-yellow_cube_constrained_front_singleThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so_follower",
"total_episodes": 51,
"total_frames": 20669,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:51"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/XixinZhang/pick-place-yellow_cube_constrained_front_single.terminal_bench_2_exp_syh_r2egym_askllm_constrained_glm_4_7_traces_jupiter_202602f9a0073exp-syh-tezos-askllm-constrainedgaia_127_laion_exp-syh-tezos-askllm-constrained_glm_4_7_traces_jupiterLe-Mat-Bulk_matgen_constrained
Le-Mat-Bulk_matgen_constrained
This repository contains preprocessed datasets from the LeMat-Bulk dataset prepared for MatterGen-style retraining.
🚀 Loading the Dataset
Since this repository is public, you can load the dataset subsets directly without authentication using the Hugging Face datasets library:
from datasets import load_dataset
# Load Default (Dataset A)
dataset_a = load_dataset("youssefsassi/Le-Mat-Bulk_matgen_constrained", name="default")
# Load… See the full description on the dataset page: https://huggingface.co/datasets/youssefsassi/Le-Mat-Bulk_matgen_constrained.terminal_bench_2_exp_syh_tezos_askllm_constrained_glm_4_7_traces_jupiter_2026025f196559terminal_bench_2_exp_syh_r2egym_askllm_constrained_glm_4_7_traces_jupiter_cleanfe7e8510exp-syh-tezos-askllm-constrained_glm_4.7_traces_jupiter_cleanedDPO_5b_llama_3600_constrained_quality_aware_bleurt