datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
drone_stl_env0000_v1_nogroundThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "uav",
"total_episodes": 500,
"total_frames": 100000,
"total_tasks": 5,
"total_videos": 1000,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 10,
"splits": {
"train": "0:500"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/Celina717/drone_stl_env0000_v1_noground.SimulCost-Bench
SimulCost-Bench
📖 Paper | 🛠️ Code | 🌐 Website | 💾 Cache (Baseline) | 💾 Cache (Full)
SimulCost is a cost-aware benchmark and toolkit for evaluating how well LLM agents tune simulation parameters under realistic computational budgets. Unlike prior evaluations that focus on correctness while implicitly treating tool usage as “free,” SimulCost explicitly measures both: (1) whether a proposed configuration meets an accuracy target and (2) how much simulation compute it consumes.The… See the full description on the dataset page: https://huggingface.co/datasets/Rose-STL-Lab/SimulCost-Bench.stl_high_complexityeval_act_stl_4cam_viewdrop_v060_test_20260923_162208This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"shape": [
6
],
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos",
"gripper.pos"… See the full description on the dataset page: https://huggingface.co/datasets/KoukiHagiwara/eval_act_stl_4cam_viewdrop_v060_test_20260923_162208.so101_cube_pick6This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": null,
"total_episodes": 10,
"total_frames": 1000,
"total_tasks": 1,
"total_videos": 10,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 50,
"splits": {
"train": "0:10"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/stlee601/so101_cube_pick6.STLBench
STLBench
STLBench is a function-level dataset for C++ STL container type recovery from stripped binaries. Each record pairs a function's ground-truth C++ source with its stripped-binary disassembly, control-flow graph (CFG), Ghidra pseudo-code, and a multi-label set of STL container types used by the function.
Anonymous release for double-blind review. Authors and project repository will be revealed after the review period.
Summary
Records: 14,884 (3,894 unique… See the full description on the dataset page: https://huggingface.co/datasets/sonia12138/STLBench.stl10-c
STL-10-C
A corrupted variant of the STL-10 test split, generated for controlled mixed-batch OOD experiments.
Dataset Description
600,000 corrupted images produced by applying imagecorruptions to STL-10's 8,000-image test split across 15 corruption types and 5 severity levels.
Each record contains:
image: PNG bytes
label: integer class label (0–9)
corruption_name: corruption type (e.g. gaussian_noise)
corruption_level: severity (1–5)
Intended Use… See the full description on the dataset page: https://huggingface.co/datasets/Kipardox/stl10-c.stl_updated
Dataset Card for stl_updated
Dataset Description
The stl_updated dataset is a large-scale collection of 3.3 million Signal Temporal Logic (STL) formulae, designed to stress-test and train models (such as Transformer encoders) on recursive understanding, semantic similarity, and syntactic complexity.
The training and test sets are augmented directly from the seed formulae introduced in Candussio (2025) and originally hosted in the base dataset saracandu/stl_formulae. These… See the full description on the dataset page: https://huggingface.co/datasets/saracandu/stl_updated.high_stldlibarray_split_func_stryker