datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
eval_smolvla_acg_baseThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so101_follower",
"total_episodes": 20,
"total_frames": 4182,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 500,
"fps": 30,
"splits": {
"train": "0:20"
},
"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/danipaez/eval_smolvla_acg_base.Sampled_LEXameval_smolvla_acg_incoThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so101_follower",
"total_episodes": 12,
"total_frames": 2440,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 500,
"fps": 30,
"splits": {
"train": "0:12"
},
"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/danipaez/eval_smolvla_acg_inco.namuwiki-extracted-acg-filtered
namuwiki-extracted-acg-filtered
ACG(Anime, Comics and Games) filtered version of heegyu/namuwiki-extracted
애니메이션, 만화, 게임 필터링 된 나무위키 덤프입니다.
How it was made?
English
Remove low effort entries: under 200 char or 3 sentences articles.
LLM-as-Judge categorization using Qwen3-Next-80B-A3B
Filtered only POI categories ("Game", "Animation", "Comic/Webtoon", "Light novels", "Vocaloid/Doujin music", "Sexual", "Insult/Internet meme", "Vtuber")
Add redirect column for… See the full description on the dataset page: https://huggingface.co/datasets/hell0ks/namuwiki-extracted-acg-filtered.noobreward-acgq-data
noobreward-acgq-data
Training and validation data of the acgq ACG aesthetic-quality reward model (model: Yidhar/noobreward-acgq; code:
GitHub Yidhar/noobreward). 120,087 images with human good/bad votes, per-dimension human scores, a Gemini teacher, WD-tagger
priors, metadata labels (artist / rating / year / score / characters) and the derived validation set.
Images (images/)
aligned-000.tar … aligned-011.tar (about 1 GB each, plain tar, 120,491 JPEG files) hold… See the full description on the dataset page: https://huggingface.co/datasets/Yidhar/noobreward-acgq-data.acg-audio-tags-and-text-generated