datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
t2v-distill-benchmark
T2V Distillation Benchmark
A benchmark comparison of Text-to-Video distillation/acceleration models based on Wan2.1-14B.
Models Compared
Model
Steps
Source
FastVideo CausalWan2.2
8
FastVideo/CausalWan2.2-I2V-A14B-Preview-Diffusers
Krea Realtime-Video
4
krea/krea-realtime-video
LightX2V CausVid
9
lightx2v/Wan2.1-T2V-14B-CausVid
NVlabs rCM 14B
4
worstcoder/rcm-Wan
Helios-Distilled
3 (pyramid 2-2-2)
BestWishYsh/Helios-Distilled… See the full description on the dataset page: https://huggingface.co/datasets/hffordata/t2v-distill-benchmark.08151422_G1CaTra2ADagger_distill_v1xGg10xGd10xLg05xLd10xOg10xOd10xB10xBg05xT004sam31-encoder-distillation-frames-sharded
SAM 3.1 encoder distillation frames
This public dataset contains 100,000 unlabeled frames in ten tar shards. Download all files under shards/, extract them into a single frames/ directory, and use manifest.json. The manifest rows already use frames/{index:06d}.jpg paths. The tar files are uncompressed because the source JPEGs are already compressed.
distill_from_human_resetThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "aloha",
"total_episodes": 202,
"total_frames": 89419,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 50,
"splits": {
"train": "0:202"
},
"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/pyrrosk/distill_from_human_reset.distill_from_buffer_resetThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "aloha",
"total_episodes": 202,
"total_frames": 90023,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 50,
"splits": {
"train": "0:202"
},
"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/pyrrosk/distill_from_buffer_reset.so101-lift-cube-residual-distillThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so101_follower",
"total_episodes": 10,
"total_frames": 779,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:10"
},
"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/igor-saprygin/so101-lift-cube-residual-distill.fr3-cube-rgb-distillation-80k
FR3 3-Cube RGB Distillation 80K
Gated LeRobot-format success-only demonstrations collected from the FR3
3-cube full-stack teacher for RGB visuomotor-policy distillation.
Summary
80,000 successful episodes
3,726,217 transitions at 10 Hz (about 103.5 hours)
320 x 180 RGB observations
Three cameras: two fixed D435 views and one wrist D405 view
Seven arm-joint actions plus gripper control in the stored action contract
Episodes stop immediately after full-stack… See the full description on the dataset page: https://huggingface.co/datasets/jaeikkim/fr3-cube-rgb-distillation-80k.